{"url": "https://arxiv.org/abs/1605.07093", "title": "Maps of Degree 1 and Lusternik--Schnirelmann Category", "abstract": "Given a map $f: M \\to N$ of degree 1 of closed manifolds. Is it true that the Lusternik--Schnirelmann category of the range of the map is not more that the category of the domain? We discuss this and some related questions.", "text": "\\section{Introduction}\r\n\r\n Below $\\cat$ denotes the (normalized) Lusternik--Schnirelmann category,~\\cite{CLOT}.\r\n \r\n \\m\r\n\\begin{quest}\\label{question}\r\nLet $M, N$, $\\dim M=\\dim N=n$, be two closed connected orientable manifolds, and let $f: M \\to N$ be a map of degree $\\pm 1$. Is it true that $\\cat M \\geq \\cat N$? \r\n\\end{quest}\r\n\r\nThere are several reasons to conjecture that the above-mentioned question has the affirmative answer for all maps $f$ of degree 1. Indeed, informally, the domain of $f$ is more ``massive'' than the range of $f$. For example, it is well known that the induced maps $f_*: H_*(M) \\to H_*(N)$ and  $f_*: \\pi_1(M) \\to \\pi_1(N)$ are surjective. \r\n\r\n\\m In~\\cite{R2} I proved some results that confirm the conjecture under some suitable hypotheses, and now people speak about the Rudyak conjecture, cf. \\cite[Open Problem 2.48]{CLOT},~ \\cite{D}, (although I prefer to state questions rather than conjectures). To date we do not know any counterexample.\r\n\r\n\\m Here we demonstrate more situations when the conjecture holds, \r\nand discuss some variations of the conjecture.\r\n\r\n\\section{Preliminaries}\r\n\r\n\\begin{df}\\rm Let $M,N, \\dim M=\\dim N=n$, be two closed connected oriented smooth manifolds, and let $[M]\\in H_n(M)=\\Z$ and $[N]\\in H_n(N)=\\Z$ be the corresponding fundamental classes. Given a map $f: M\\to N$, we define the {\\em degree} of a map $f:M \\to N$ as the number $\\deg f\\in \\Z$ such that $f_*[M]=(\\deg f) [N]$. \r\n\\end{df}\r\n\r\nLet $E$ be a commutative ring spectrum (cohomology theory).\r\n\r\n\\begin{lem}\\label{mono} If $M$ is $E$-orientable and $f: M \\to N$ is a map of degree $1$ then $f_*\\colon E_*(M)\\to E_*(N)$ is an epimorphism and $f^*\\colon E^*(N)\\to E^*(M)$ is a monomorphism.  \r\n\\end{lem}\r\n\r\n\\p See~\\cite[Theorem V.2.13]{R1}, cf. also Dyer~\\cite{Dyer}.\r\n\\qed\r\n\r\n\\begin{remark}\\label{r:local}\\rm An analog of \\lemref{mono} holds for cohomology with local coefficients.  Let $\\A$ be a local coefficient system of abelian groups on $N$, and let $f^*(\\A)$ be the induced coefficient system. Then $f_*:H_*(M;f^*\\A)\\to H_*(N;\\A)$ is a split epimorphism and $f^*\\colon H^*(N;\\A)\\to H^*(M;f^*\\A)$ is a split monomorphism. The proofs are based on Poincar\\'e duality with local coefficients and the equality $f_*(f^*x\\frown y)=x\\frown f_* y$ for  $x\\in H^*(N;\\A), y\\in H_*(M;f^*\\A)$. See e.g.~\\cite{B}.\r\n\\end{remark}\r\n\r\n\\begin{df}\\rm  Given a CW space $X$, the {\\em Lusternik--Schnirelmann category} $\\cat X$ of $X$ is the least integer $m$ such that there exists an open covering $\\{A_0, A_1,\\ldots, A_m \\}$ with each $A_i$ contractible in $X$ (not necessary in itself). If no such covering exist, we put  $\\cat X=\\infty$. \r\n\\end{df}\r\n\r\n Note the inequality $\\cat X \\leq \\dim X$ for $X$ connected.\r\n\r\n\\m In future we abbreviate Lusternik--Schnirelmann to LS. A good source for LS theory is~\\cite{CLOT}.\r\n\r\n\\m Given a closed smooth manifold $M$ and a smooth function $f: M \\to \\R$, the number of critical points of $f$ can't be less than $\\cat M$,~\\cite{LS1, LS2}. This result turned out to be the starting point of LS theory. Currently, the LS theory is a broad area of intensive topological research.\r\n\r\n\\m Let $X$ be a path connected space. Take a point $x_0\\in X$, set \r\n$$\r\nPX=P(X,x_0)=\\{\\omega\\in X^{[0,1]} \\bigm| \\omega(0)=x_0\\}\r\n$$\r\nand consider the fibration $p: PX \\to X, p(\\omega)=\\omega(1)$. Let $p_k=p_{k}^X: P_k(X)\\to X$ be the $k$-fold fiberwise join $PX*_X\\cdots *_X* PX \\to X$. According to the Ganea--\\v Svarc theorem, \\cite{CLOT, S},  the inequality $\\cat X<k$ holds if and only if the fibration $p_k: P_k(X)\\to X$ has a section. In other words, the number $\\cat X$ is the least $k$ such that the fibration $p_{k+1}: P_{k+1}(X)\\to X$ has a section.\r\n\r\n\\section{Approximations}\r\n\r\nRecall (see the Introduction) that we discuss whether the existence of a map $f: M \\to N$ of degree 1 implies the inequality $\\cat M \\geq \\cat N$. Here we present two  results appearing when we approximate the LS category by the cup-length and Toomer invariant. \r\n\r\n\\m Recall the following cup-length estimate of LS category. Let $E$ be a commutative ring spectrum (cohomology theory). The cup-length of $X$ with respect to $E$ is the number \r\n\\[\r\n\\cl_E(X):=\\sup \\{m\\bigm| u_1\\smallsmile\\cdots \\smallsmile u_m \\neq 0 \\text{ where } u_i\\in\\wt E^*(X)\\}.\r\n\\]\r\nThe well-known cup-length theorem~\\cite{CLOT} asserts that $\\cl_E(X)\\leq \\cat X$.\r\n\r\n\\m We give another estimate of  LS category. \r\n\r\n\\begin{df}\\label{d:toomer} \\rm Define the (cohomological) Toomer invariant \r\n\\[\r\ne_E^*(X)=\\sup\\{k\\bigm | \\ker \\{p_k^*: E^*(X))\\to E^*(P_k(X))\\}\\neq 0\\}.\r\n\\] \r\n\\end{df}\r\n\r\nNote the decreasing sequence $\\cdots \\supset \\ker (p_k^*)\\supset \\ker (p_{k+1}^*)\r\n\\supset \\cdots$. Moreover,\r\n\\[\r\ne_E^*(X)=\\inf\\{k\\bigm | \\ker \\{p_k^*: E^*(X)\\to E^*(P_k(X))\\}= 0\\}-1.\r\n\\] \r\n\r\n\\m In the definition of the Toomer invariant, $E$ does not need to be a ring spectrum, it can be an arbitrary spectrum.\r\n \r\n\\begin{prop}\\label{p:doubleineq}\r\n We have $\\cl_E(X)\\leq e_E^*(X)\\leq \\cat X$.\r\n\\end{prop}\r\n\r\n\\p First, $p_m=p_{m}^X$ has a section for $m> \\cat X$, and so $\\ker p_{m}^*=0$ for all $m>\\cat X$. Hence, $e_E^*(X)\\leq \\cat X$. Now, we put $\\cl_E(X)=l$, $e_E^*(X)=k$ and prove that $l\\leq k$. Take $u_1, \\ldots, u_l\\in \\wt E^*(X)$ such that $u_1\\smallsmile \\cdots \\smallsmile u_l\\neq 0$. Then, since $p_{k+1}^*$ is monic, we conclude that $p_{k+1}^*(u_1\\smallsmile \\cdots \\smallsmile u_l)\\neq 0$. In other words, \r\n\\[(p_{k+1}^*u_1)\\smallsmile \\cdots \\smallsmile  (p_{k+1}^* u_l)\\neq 0\r\n\\]\r\n in $P_{k+1}X$. Hence, $\\cat P_{k+1}(X)\\geq l$ because of the cup-length theorem. It remains to note that $\\cat P_{k+1}(X)\\leq k$ for all $k$, see~\\cite[Prop. 1.5(ii)]{R2} or~\\cite{CLOT}.\r\n \\qed\r\n\r\n\\begin{prop}\\label{p:ineq} Let $M$ be a closed connected $E$-orientable manifold, and let $f: M \\to N$ be a map of degree $\\pm 1$. Then $\\cl_E(M)\\geq \\cl_E(N)$ and $e^*_E(M)\\geq e^*_E(N)$.\r\n\\end{prop}\r\n\r\n\\p Put $\\cl_E(N)=l$ and take $u_1, \\ldots u_l\\in \\wt E^*(N)$ such that $u_1\\smallsmile \\cdots \\smallsmile u_l\\neq 0$. Then $f^*(u_1\\smallsmile \\cdots \\smallsmile u_l)\\neq 0$ by \\ref{mono}. So, $f^*(u_1)\\smallsmile \\cdots \\smallsmile f^*(u_l)\\neq 0$ in $M$, and thus $\\cl_E(M)\\geq l$.\r\n\r\n\\m Now, let $e^*_E(M)=k$. Then $(p_i^M)^*$ is a monomorphism for $i>k$. Consider the commutative  diagram\r\n\\[\r\n\\CD\r\nP_iM @>P_if>> P_i N\\\\\r\n@Vp_i^M VV @Vp_i^N VV\\\\\r\nM@>f>> N.\r\n\\endCD\r\n\\]\r\nand note that, by \\ref{mono}, $f^*((p_i^M)^*)$ is monic for $i>k$. Hence, because of commutativity of the diagram, $(p_i^N)^*$ is monic for $i>k$. Thus, $e^*_E(N)\\leq k$. \r\n\\qed\r\n\r\n\\begin{cor}\\label{c:cat=cl}\r\nLet $M$ be a closed connected $E$-orientable manifold, and let $f: M \\to N$ be a map of degree $\\pm 1$.\r\n\r\n{\\rm(i)} Suppose that $e^*_E(N)=\\cat N$. Then $\\cat M\\geq \\cat N$.\r\n\r\n {\\rm(ii)} Suppose that $\\cl_E(N)=\\cat N$. Then $\\cat M\\geq \\cat N$. \r\n\\end{cor}\r\n\r\n\\p (i) We have $\\cat M \\geq e^*_E(M)\\geq e^*_E(N)=\\cat N$, the second inequality following from \\propref{p:ineq}. \r\n\r\n(ii) Because of (i), it suffices to prove that  $e^*_E(N)=\\cat N$. But this holds since $\\cl_E(N)\\leq e^*_E(N)\\leq \\cat N$.\r\n\\qed\r\n\r\n\r\n\r\n\r\n\\section{Low-Dimensional Manifolds}\\label{s:low} \r\n\r\nWe prove that for $n\\leq 4, \\cat M^n \\geq \\cat N^n$ provided that there exists a map $f: M \\to N$ of degree 1. The inequality holds trivially for $n=1$. \r\n\r\n\\m The case $n=2$ is also simple. Denote by $g(X)$ the genus of a closed connected orientable surface $X$. Then $g(M)\\geq g(N)$ because of surjectivity of $f_*: H_2(M) \\to  H_2(N)$, see~\\ref{mono}.  Furthermore, $\\cat X=1$ if $g(X)=0$ ($ X=S^2)$ and $\\cat X=2$ for $g>1$. \r\n\r\n\\m The case $n=3$ is considered in~\\cite[Corollary 1.3]{OR}.\r\n\r\n\\m The case $n=4$.\r\nFirst, if $\\cat N=4$ then $\\cat N=\\cat M$ by~\\cite[Corollary 3.6(ii)]{R2}. \r\n\r\n\\m\r\nNext we consider $\\cat N=3$ and prove that $\\cat M \\geq 3$. By way of contradiction, assume that $\\cat M\\leq 2$, and hence the group $\\pi_1(M)$ is free by~\\cite{DKR}. \r\nNow, $\\pi_1(N)$ is free since $\\deg f=1$, see~\\cite{DR}. But then $N$ has a CW decomposition whose 3-skeleton is a wedge of spheres,~\\cite{Hil}, and hence $\\cat N\\leq 2$, a contradiction. Finally, the case $\\cat M=1<2=\\cat N$ is impossible for trivial reasons ($M$ should be a homotopy sphere, and therefore $N$ should be a homotopy sphere).\r\n\r\n\\section{Some exemplifications}\r\n\r\nThe following result is a weak version of \\cite[Theorem 3.6(i)]{R2}.\r\n\r\n\\begin{thm} \\label{t:main}\r\nLet $M, N$ be two smooth closed connected stably parallelizable manifolds, and assume that there exists a map $f: M \\to N$ of degree $\\pm 1$. If $N$ is $(q-1)$-connected and $\\dim N\\leq 2q\\cat N-4$ then $\\cat M \\geq \\cat N$.\r\n\\qed\r\n\\end{thm}\r\n\r\n\\begin{remark}\\rm\r\n In~\\cite[Theorem 3.6(i)]{R2} we use~\\cite[Corollary 3.3(i)]{R2} where, in turn, we require $\\dim N \\geq 4$. However, the case $\\dim N \\leq 4$ is covered by Section~\\ref{s:low}.\r\n \\end{remark}\r\n\r\n\\begin{thm} \\label{t:torus}\r\nLet $T^k$ denote the $k$-dimensional torus. Let $M, N$ be two smooth closed connected stably parallelizable  manifolds, and assume that there exists a map $f: M \\to N$ of degree $\\pm 1$. Then there exists $k$ such that $\\cat (M \\ts T^k)\\geq \\cat (N\\ts T^k)$. \r\n\\end{thm}\r\n\r\n\\p Put $\\dim M=n$ and note that $\\cat (T^k\\ts M)\\geq \\cat T^k=k$. Now, if $k\\geq n+4$ then\r\n\\[\r\n2\\cat M-4\\geq 2k-4 \\geq k+n,\r\n\\]\r\nand we are done by \\theoref{t:main}.\r\n\\qed\r\n\r\n\\m Another example. Consider the exceptional Lie group $G_2$. Recall that $\\dim G_2=14$. Note that  $G_2$ is parallelizable being a Lie group.\r\n\r\n\\begin{prop} Let $M$ be a stably parallelizable 14-dimensional closed manifold that admits a map $f: M \\to G_2$ of degree $\\pm 1$. Then $\\cat M \\geq \\cat G_2$.\r\n\\end{prop}\r\n\r\n\\p The group $G_2$ is 2-connected and $\\cat G_2=4$,~\\cite{IM}. Now the result follows from \\theoref{t:main} with $N=G_2$ and $q=3$ because $14=\\dim G_2\\leq 2q\\cat G_2-4=20$.\r\n\\qed\r\n\r\n\\m Let $SO_n$ denote the special orthogonal group, i.e., the group of the orthogonal $n\\times n$-matrices of determinant 1. Recall that  $\\dim SO_n=n(n-1)/2$.\r\n\r\n\\begin{thm} \\label{t:son}\r\nLet $M$ be a closed connected smooth  manifold, and let $f: M \\to SO_n$ be a map of degree $\\pm 1$. Then $\\cat M \\geq \\cat(SO_n)$ for $n\\leq 9$.\r\n\\end{thm}\r\n\r\n\r\n\\p We apply \\corref{c:cat=cl} for the case $E=H\\Z/2$, i.e., to arbitrary closed connected manifolds.  Below $\\cl$ denotes $\\cl_{\\Z/2}$. Because of \\corref{c:cat=cl} it suffices to prove that $\\cat SO_n=\\cl(SO_n)$ for $n\\leq 9$. Recall that $H^*(SO_n;\\Z/2)$ is the polynomial algebra on generators $b_i$ of odd degree $i<n$, truncated by the relations ${b_i^{p_i}}=0$ where $p_i$ is the smallest power of 2 such that ${b_i^{p_i}}$ has degree $\\geq n$,~\\cite{H}. In other words,\r\n\\[\r\nH^*(SO_n;\\Z/2)=\\Z/2[b_1, \\ldots, b_k, \\ldots]/(b_1^{p_1},  \\ldots, b_k^{p_k}, \\ldots).\r\n\\]\r\nNote  that $p_k=1$ for $2k-1>n$, and so $H^*(SO_n)$ is really a truncated polynomial ring (not a formal power series ring). Hence, \r\n \\[\r\n \\cl(SO_n)=(p_1-1)+\\ldots +(p_k-1)+\\ldots\r\n \\]\r\n and the sum on the right is finite because $p_k=1$ for all but finitely many $k$'s.\r\n \r\n\\m For sake of simplicity,  we use the notation $S_n$ for  $H^*(SO_n;\\Z/2)$. We have:\r\n\r\n$\\dim SO_3=3$, $\\cl(SO_3)=3 \\text{ because } S_3=\\Z/2[b_1]/(b_1^4)$.\\\\\r\n$\\dim SO_4=6$, $\\cl(SO_4)=4 \\text{ because } S_4=\\Z/2[b_1,b_3]/(b_1^4,  b_3^2)$.\\\\\r\n$\\dim SO_5=10$, $\\cl(SO_5)=8 \\text{ because } S_5=\\Z/2[b_1,b_3]/(b_1^8,  b_3^2)$.\\\\\r\n$\\dim SO_6=15$, $\\cl(SO_6)=9 \\text{ because } S_6=\\Z/2[b_1,b_3, b_5]/(b_1^8,  b_3^2, b_5^2)$.\\\\\r\n$\\dim SO_7=21$, $\\cl(SO_7)=11\\text{ because } S_7=\\Z/2[b_1,b_3, b_5]/(b_1^8,  b_3^4, b_5^2)$.\\\\\r\n$\\dim SO_8=28$, $\\cl(SO_8)=12 \\text{ because } S_8=\\Z/2[b_1,b_3, b_5, b_7]/(b_1^8,  b_3^4, b_5^2, b_7^2)$.\\\\\r\n$\\dim SO_9=36$, $\\cl(SO_9)=20 \\text{ because } S_9=\\Z/2[b_1,b_3, b_5, b_7]/(b_1^{16},  b_3^4, b_5^2, b_7^2)$.\\\\\r\n\r\nThe values of $\\cat SO_n, n\\leq 9$ are calculated in~\\cite{IMN}, see also~\\cite{I2}. Compare these values with the above-noted values of $\\cl(SO_n)$ and see that $\\cat SO_n=\\cl(SO_n)$ for $n\\leq 9$. \r\n\\qed\r\n\r\n\r\n\\begin{remarks}\\rm  1. The anonymous referee noticed that, probably, the method of \\theoref{t:son} can also be applied to other Lie groups ($U_n, SU_n$, etc.). This is indeed true, but we do not develop these things here. \r\n\r\n \\m 2. In above-mentioned \\theoref{t:main} we can relax the assumption on $M$ by requiring that the normal bundle of $M$ be stably fiber homotopy trivial, i.e., that $M$ is $S$-orientable where $S$ is the sphere spectrum. However, we can't provide the same weakening for $N$ because our proof in~\\cite{R2} uses surgeries on $N$. \r\n\\end{remarks}\r\n\r\n\\section{Theme and Variations: Other Numerical Invariants Similar to LS Category}\r\n\r\nLet $\\crit X$ denote the minimum number of critical points of a smooth function $f:X \\to \\R$ on the closed smooth manifold $X$, and let $\\ballcat (X)$ be the minimum $m\\in \\N$ such that there is a covering of $X$ by $m+1$ smooth open balls. It is known that $\\cat X \\leq \\ballcat(X)\\leq \\crit M-1$, see~\\cite{CLOT}. Note that there are examples with $\\cat M<\\ballcat(M)$. Indeed, there are examples of manifolds $M$ such that $\\cat M=\\cat (M\\setminus \\pt)$,~\\cite{I1}, while $\\cat (M\\setminus \\pt)=\\cat M-1$ whenever $\\cat M=\\ballcat(M)$. On the other hand, there are no known examples with $\\ballcat M +1<\\crit M$.\r\n\r\n \\m Now,  we can pose an open question whether $\\ballcat M\\geq \\ballcat N$ and $\\crit M \\geq \\crit N$ provided there exists a map $f: M \\to N$ of degree 1.\r\n \r\n\\m We can also consider the number $\\crit^*(X)$, that is, the minimum number of {\\em nongenerate} critical points of a smooth function $f:X \\to \\R$ on the closed smooth manifold $X$. There is a big difference between $\\crit X$ and $\\crit^*X$. For example, if $S_g$ is a surface of genus $g\\geq 1$ then $\\crit S_g=3$ while $\\crit^*S_g=2g$. So, we can ask if $\\crit^* M \\geq \\crit^* N$ provided there exists a map $f: M \\to N$ of degree~1. One of the lower bounds of $\\crit^*(X)$ is the sum of Betti numbers $\\SB(X)$, the inequality $\\SB(X)\\leq\\crit^*(X)$ being a direct corollary of Morse theory, \\cite{M}, and we can regard $\\SB(X)$ as an approximation of $\\crit^*X$. Now, if $f: M \\to N$ is a map of degree~1 then the inequality $\\SB(M)\\geq \\SB(N)$ follows from \\ref{mono}.\r\n\r\n\\m If $M,N$ are closed simply-connected manifolds of dimension $\\geq 6$ and there exists a map $M\\to N$ of degree~1 then $\\crit^*(M)\\geq \\crit^*(N)$. Indeed, for every Morse function $h: X\\to  \\R$ on a closed connected smooth manifold $X$ we have the Morse inequalities \r\n\\[\r\nm_{\\gl}\\geq r_{\\gl}+t_{\\gl}+t_{\\gl-1}\r\n\\]\r\n where $m_{\\gl}$ is the number of critical points of index $\\gl$ of $h$ and $r_{\\gl}, t_\\lambda$ are the rank and the torsion rank of $H_{\\gl}(X)$, respectively.  Now, if $X$ is a closed simply connected manifold with $\\dim M \\geq 6 $ then $X$ possesses a Morse function for which the above-mentioned Morse inequalities turn out to be equalities. This is a well-known Smale Theorem~\\cite{Sm}. Now, the inequalities $r_{\\gl}(M)\\geq r_{\\gl}(N)$ and $t_{\\gl}(M) \\geq t_{\\gl}(N)$ follow from \\ref{r:local}, and we are done.\r\n\\qed\r\n\r\n\\m {\\bf Acknowledgments:} The work was partially supported by a grant from the Simons Foundation (\\#209424 to Yuli Rudyak). I am very grateful to the anonymous referee who assisted a lot in my work with \\corref{c:cat=cl} and \\theoref{t:son},  and made other valuable suggestions that improved the paper.\r\n\r\n", "meta": {"timestamp": "2016-09-27T02:00:55", "yymm": "1605", "arxiv_id": "1605.07093", "language": "en", "url": "https://arxiv.org/abs/1605.07093", "abstract": "Given a map $f: M \\to N$ of degree 1 of closed manifolds. Is it true that the Lusternik--Schnirelmann category of the range of the map is not more that the category of the domain? We discuss this and some related questions.", "subjects": "Algebraic Topology (math.AT)", "title": "Maps of Degree 1 and Lusternik--Schnirelmann Category", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9835969655605173, "lm_q2_score": 0.8633916205190225, "lm_q1q2_score": 0.8492293780328882}}
{"url": "https://arxiv.org/abs/1810.00492", "title": "Chords of an ellipse, Lucas polynomials, and cubic equations", "abstract": "A beautiful theorem of Thomas Price links the Fibonacci numbers and the Lucas polynomials to the plane geometry of an ellipse, generalizing a classic problem about circles. We give a brief history of the circle problem, an account of Price's ellipse proof, and a reorganized proof, with some new ideas, designed to situate the result within a dense web of connections to classical mathematics. It is inspired by Cardano's solution of the cubic equation and Newton's theorem on power sums, and yields an interpretation of generalized Lucas polynomials in terms of the theory of symmetric polynomials. We also develop additional connections that surface along the way; e.g., we give a parallel interpretation of generalized Fibonacci polynomials, and we show that Cardano's method can be used write down the roots of the Lucas polynomials.", "text": "\\section{Introduction}\n\nA classic problem\ninstructs the reader to mark off $n$-equally spaced points around on a unit circle, draw the chords connecting one of the points to the remaining $n-1$ others, and find the product of the lengths.\n\n\\begin{figure}[htbp]\\label{fig:circlechords}\n\\begin{center}\n\\resizebox{2in}{!}{\\includegraphics{CircleDiagonals}}\n\\caption{Chords of a Circle}\n\\label{diagonals}\n\\end{center}\n\\end{figure}\n\nAlthough most or all of the lengths are irrational, the product is exactly $n$. An elegant solution (\\cite{galovich},\n \\cite{price1}, \\cite{silva}) involves the $n$th roots of unity $1, \\zeta, \\zeta^2, \\ldots, \\zeta^{n-1}$, which are equally spaced around the unit circle in the complex plane. Connecting $1$ to each of the $n - 1$ others and multiplying the lengths of the segments gives the product:\n$$\\Pi_n = |1 - \\zeta | |1 - \\zeta^2| \\cdots |1 - \\zeta^{n-1}| = |(1 - \\zeta)(1 - \\zeta^2) \\cdots (1 - \\zeta^{n-1})|.$$\n\nThis is the absolute value of the polynomial $$\\Pi_n(z) = (z - \\zeta)(z - \\zeta^2) \\cdots (z - \\zeta^{n-1})$$ evaluated at $z=1$.\n\nThe $n$th roots of unity are, by definition, the roots of $z^n-1$.  Now $\\Pi_n$ has as roots all roots of unity except 1; it follows that $\\Pi_n(z) = \\frac{z^n - 1}{z - 1}=z^{n-1}+\\dots+z+1$.  Thus the answer desired is $|\\Pi_n(1)| = n$.\n\nThis problem has an interesting history, of which we discuss a few highlights below in section \\ref{sec:circlehistory}.\n\nIn \\cite{price1}, Thomas E. Price considered the following generalization. Scale Figure \\ref{fig:circlechords} horizontally and/or vertically. It becomes an ellipse. What happens to the product of the chord lengths? An elegant explicit formula is given in \\cite{price1}, which we re-derive below (Proposition \\ref{prop:price}).\n\n\\begin{figure}[htbp]\n\\begin{center}\n\n\\resizebox{2in}{!}{\\includegraphics{StretchedDiagonals}}\n\n\\caption{Stretched Chords}\n\\label{fig:ellipsechords}\n\\end{center}\n\\end{figure}\n\nShowcasing the power of his results, Price in \\cite{price2} considered the special case of a vertical stretch factor of $\\sqrt{5}$ (see Figure \\ref{fig:ellipsechords}). In this situation, the product of the chord lengths is given in Table \\ref{tbl:ellipseproducts}. The answer is remarkable.\n\n\\begin{table}\n\\begin{center}\n\\begin{tabular}{c|c}\n$n$ & product\\\\ \\hline\n2 & $2 = 2\\cdot 1$\\\\\n3 & $6 = 3\\cdot 2$\\\\\n4 &  $12 = 4\\cdot 3$\\\\\n5 &  $25 = 5\\cdot 5$\\\\\n6 & $48 = 6\\cdot 8$\\\\\n7 & $91= 7\\cdot 13$\\\\\n\\vdots & \\vdots \\\\\n$n$ & $nF_n$\n\\end{tabular}\n\\end{center}\n\\caption{Products of stretched chords. The symbol $F_n$ refers to the $n$th Fibonacci number.}\\label{tbl:ellipseproducts}\n\\end{table}\n\nOur purpose in this note is to offer an alternative proof of this beautiful and insufficiently-well-known result of Price. We recover the result via a conceptually transparent argument that highlights several gems of classical mathematics. It is inspired by the classical Cardano formula for the solution of a reduced cubic equation by radicals, and makes use of a formula of Newton expressing the elegant relation between elementary symmetric polynomials and power sums. In the situation of interest, Newton's formula becomes the two-term linear recurrence that characterizes the generalized Lucas polynomials. Thus, the generalized Lucas polynomials also generalize the reduced cubic. The method provides an interpretation of the generalized Lucas polynomials that makes their Binet formula fall out as an immediate consequence. We give an analogous interpretation of the generalized Fibonacci polynomials that justifies their Binet formula as well.\n\nWe integrate our proof of Price's result with exposition of the classical mathematics involved. Thus, we aim for a completely self-contained treatment. We include short proofs of Cardano's and Newton's formulas, as well as all the facts about Lucas and Fibonacci polynomials that are needed.\n\nThe structure of the paper is as follows. In section \\ref{sec:cardano}, we give a short proof of Cardano's formula for solving a reduced cubic polynomial in radicals, due to Solomon, and show that the polynomial that serves for the ellipse the role played above by $z^n-1$ for the circle, which we call $\\Omega_n(z)$, generalizes the reduced cubic. We show that $\\Omega_n(z)$ has a natural interpretation in terms of the fundamental theorem on symmetric polynomials (FTSP). In section \\ref{sec:newtonbinet}, we sketch a proof of Newton's theorem on power sums, and then use it to derive a recursive formula for the key part of $\\Omega_n(z)$. We observe that this recurrence is actually none other than the recurrence defining the generalized Lucas polynomials, up to a sign change, and therefore $\\Omega_n(z)$ is expressed in terms of the generalized Lucas polynomials. The Binet formula for the generalized Lucas polynomial now falls out as a consequence of the interpretation in terms of the FTSP. In section \\ref{sec:solution}, we use known facts about generalized Lucas and Fibonacci polynomials to derive Price's beautiful formula \\ref{prop:price} for the product of the stretched elliptical chords. In section \\ref{sec:solnbyradicals}, we bring the inquiry full-circle by giving the solution of $\\Omega_n(z)$ by radicals, which generalizes the Cardano formula, and recovers the known formula for the roots of the Lucas polynomial as a special case. The rest of the paper consists of commentary and loose ends. Price's starting point, the circle problem described above, has an interesting history, and we give some highlights in section \\ref{sec:circlehistory}. In section \\ref{sec:furtherremarks}, in the name of self-containment, we offer proofs for the facts we used in section \\ref{sec:solution}, including a proof of the Binet formula for generalized Fibonacci polynomials along the lines of what was done for the generalized Lucas polynomials in section \\ref{sec:newtonbinet}. We also discuss the relation between this and Price's work.\n\nThroughout, $\\zeta$ refers to a primitive $n$th root of unity, which can be taken to be $e^{2\\pi i / n}$. We prefer it to $\\zeta_n$ to avoid visual clutter, but the notation conceals the dependence of $\\zeta$ on $n$. Hopefully no confusion will result.\n\n\\section{A generalization of Cardano's reduced cubic}\\label{sec:cardano}\n\nSince the product of chord lengths for the circle is so elegantly found via the polynomial $z^n-1$ whose roots $1,\\zeta, \\zeta^2,\\dots$ describe the points of interest in the complex plane, the essential challenge in the generalized elliptical version of the problem is to describe the polynomial whose roots are the horizontally and vertically scaled images of these points. As Price observed in \\cite{price1}, these image points have the form\n\\begin{equation}\\label{eq:roots}\n\\zeta^ja + \\zeta^{-j}b,\n\\end{equation}\nwhere $a> |b|\\geq 0$, and $a+b$, $a-b$ are the lengths of the ellipse's horizontal and vertical axes. The original unit circle of Figure \\ref{fig:circlechords} is the case $a=1$, $b=0$. In the case where the circle is vertically scaled by $\\sqrt{5}$, as in Figure \\ref{fig:ellipsechords}, we have $a=\\phi = (1+\\sqrt{5})/2$, the golden ratio, and $b=\\hat\\phi = (1-\\sqrt{5})/2$, its algebraic conjugate.\n\nOur solution begins with the observation that, by the classical Cardano formula, the roots of a {\\em reduced cubic equation}\n\\begin{equation}\\label{eq:cubic}\nz^3 + pz + q = 0\n\\end{equation}\nalso have the form \\eqref{eq:roots}, where $n=3$, and\n\\begin{equation}\\label{eq:aandb}\na,b = \\sqrt[3]{-\\frac{q}{2}\\pm\\sqrt{\\frac{q^2}{4}+\\frac{p^3}{27}}}.\n\\end{equation}\nThe  cube roots must be chosen so that $ab = -p/3$. (In this context the word ``reduced\" refers to the fact that \\eqref{eq:cubic} is monic with no quadratic term.)\n\nThe following interpretation of Cardano's formula\nis given by Ronald Solomon in \\cite[pp.~50--51]{solomon}. The identity\n\\begin{equation}\\label{eq:identity}\n(a+b)^3 - 3ab(a+b) - (a^3 + b^3) = 0\n\\end{equation}\nholds in any field (indeed, in any commutative ring). This identity becomes the reduced cubic \\eqref{eq:cubic} if $z=a+b$, $p = -3ab$, and $q=-(a^3 + b^3)$. Thus, finding a $z$ satisfying \\eqref{eq:cubic} can be achieved by finding $a$ and $b$ satisfying $ab = -p/3$ and $a^3 + b^3 = -q$, whereupon $z=a+b$ is the desired solution. This in turn can be accomplished by noting that $ab = -p/3$ implies $a^3b^3 = -p^3/27$; thus $a^3,b^3$ are quantities whose sum and product are known. It follows that $a^3,b^3$ solve the known quadratic equation $X^2 + qX - p^3/27=0$ (the {\\em resolvent quadratic} of \\eqref{eq:cubic}). Applying the quadratic formula to this equation and taking cube roots yields \\eqref{eq:aandb}.\n\nBecause we have a choice of cube roots for $a,b$ but must preserve the known relation $ab = -p/3$, we can twist $a$ by a factor of $\\zeta$, but not without twisting $b$ by $\\zeta^{-1}$. This is why the three solutions to \\eqref{eq:cubic} have the form \\eqref{eq:roots}. \n\nWe seek a generalization of \\eqref{eq:identity} for $n>3$ with the same property, that $a$ can be replaced by $\\zeta a$ but not without replacing $b$ by $\\zeta^{-1}b$. \n\nBy moving the final term to the right side, the identity \\eqref{eq:identity} expresses the {\\em power sum} $a^3 + b^3$ in terms of the {\\em elementary symmetric polynomials} $a+b$ and $ab$. By the fundamental theorem on symmetric polynomials, higher power sums $a^n+b^n$ can also be expressed in terms of $a+b$ and $ab$, since they are symmetric in $a,b$; in fact, the expression can be taken, in a unique way, to be an integer polynomial. \n\n\\begin{definition}\\label{def:L}\nFor each $n$, let $L_n(X,Y)$ denote the unique integer polynomial such that\\begin{equation}\\label{eq:higheridentity}\nL_n(a+b,ab) - (a^n + b^n) = 0\n\\end{equation} is an identity. \n\\end{definition}\n\nWe will see below that the polynomial $L_n(X,Y)$ just defined is actually a familiar mathematical object, although it is not usually defined in this way.\n\nSince $a^n+b^n$ is homogeneous of degree $n$ in the indeterminates $a,b$, the expression $L_n(a+b,ab)$ must also be homogeneous of degree $n$. Thus $L_n(X,Y)$ is homogeneous if $X$ and $Y$ are given the weights $1$ and $2$ respectively. Since $a^n+b^n$ is not divisible by $ab$, $L_n(X,Y)$ cannot be divisible by $Y$. Thus it must have a term that does not include $Y$, and it follows that $L_n(X,Y)$ is of degree $n$ in $X$. (It is not hard to see, and we will show below,  that it is actually \\textit{monic} of degree $n$ in $X$.)\n\nThis is the desired identity generalizing \\eqref{eq:identity}: if one makes the substitution $a \\mapsto \\zeta a$, $b\\mapsto \\zeta^{-1}b$, both $ab$ and $a^n+b^n$ are left unchanged. Thus \\eqref{eq:higheridentity} becomes $L_n(\\zeta a + \\zeta^{-1}b, ab) - (a^n + b^n) = 0$. Repeating this substitution $n-2$ more times, we conclude that all the desired numbers \\eqref{eq:roots} are roots of $L_n(z,ab) - (a^n+b^n)$; bearing in mind the above remark about the degree of $L_n(X,Y)$ in $X$, we have\n\n\\begin{prop}\\label{prop:thepoly}\nThe roots of the polynomial\n\\[\n\\Omega_n(z) := L_n(z,ab) - (a^n + b^n),\n\\]\nare the $n$ numbers \\eqref{eq:roots} for $j=0,1,\\dots,n-1$, where $L_n$ is the polynomial of Definition \\ref{def:L}.\\qed\n\\end{prop}\n\nBy \\eqref{eq:higheridentity}, $\\Omega_n(z)$ can also be written $L_n(z,ab) - L_n(a+b,ab)$, implying a certain symmetry between the constant (in $z$) term and the rest. To avoid notational clutter, we have suppressed the dependence of $\\Omega_n(z)$ on $a$ and $b$.\n\n\n\n\\section{Newton's and Binet's formulas}\\label{sec:newtonbinet}\n\nTo use proposition \\ref{prop:thepoly} to find the product of the stretched elliptical chords as in figure \\ref{fig:ellipsechords}, we need more information about $\\Omega_n(z)$, or, equivalently, $L_n(X,Y)$. A classical result of Newton delivers a recursive formula for $L_n(X,Y)$.\n\n{\\em Newton's theorem} states that, for any $m$ indeterminates $a_1,\\dots,a_m$, the {\\em power sums} $s_1 = \\sum_j a_j$, $s_2 = \\sum_j a_j^2$, $\\dots$, and the {\\em elementary symmetric polynomials} $\\sigma_1 = \\sum_j a_j$, $\\sigma_2 = \\sum_{j<k} a_ja_k$, $\\sigma_3 = \\sum_{j<k<\\ell} a_{j}a_{k}a_{\\ell}$, $\\dots$, obey the following relation for any natural number $n$:\n\\begin{equation}\\label{eq:newton}\ns_n - s_{n-1}\\sigma_1 + s_{n-2}\\sigma_2 - \\dots \\pm s_1\\sigma_{n-1} \\mp n\\sigma_n = 0.\n\\end{equation}\nThe proof is a beautiful computation. The power sum $s_n$ is, of course, the sum of all terms of the form $a_j^n$. The next term of \\eqref{eq:newton}, $s_{n-1}\\sigma_1$, cancels all of these terms, but introduces new ones of the form $a_j^{n-1}a_k$. Then $s_{n-2}\\sigma_2$ cancels these, but introduces new terms $a_j^{n-2}a_ka_\\ell$, and so on, until finally $n\\sigma_n$ cancels everything remaining. The relation holds even when $n>m$, via the convention that $\\sigma_j = 0$ when $j>m$, in which case, only the first $m+1$ terms are nonzero.\n\nIn the situation of interest to us, we have $m=2$ unknowns. Thus \\eqref{eq:newton} reduces to\n\\begin{equation}\\label{eq:newton2}\n(a^n + b^n) - (a^{n-1}+b^{n-1})(a+b) + (a^{n-2}+b^{n-2})(ab) = 0\n\\end{equation}\nfor any $n\\geq 3$. In fact, since when $n=2$ we have $a^{n-2}+b^{n-2}=2=n$, it actually holds for all $n\\geq 2$, which can of course also be checked directly. Making the substitution $a^j + b^j\\mapsto L_j(a+b,ab)$ (for $j=n, n-1, n-2$) followed by $a+b \\mapsto X$ and $ab \\mapsto Y$, we obtain\n\\[\nL_n(X,Y) - XL_{n-1}(X,Y) + Y L_{n-2}(X,Y) = 0,\n\\] or\n\\begin{equation}\\label{eq:lucasrec}\nL_n(X,Y) = X L_{n-1}(X,Y) - Y L_{n-2}(X,Y),\n\\end{equation}\nvalid for all $n\\geq 2$. This linear recurrence characterizes $L_n(X,Y)$ for all $n$ once $L_0(X,Y)$ and $L_1(X,Y)$ are known. Since $L_0(a+b,ab) = a^0 + b^0 =2$, we have $L_0(X,Y) = 2$, and since $L_1(a+b,ab) = a^1 + b^1 = a+b$, we have $L_1(X,Y) = X$. It now follows from \\eqref{eq:lucasrec} by induction on $n$ that $L_n(X,Y)$ is monic in $X$, as was claimed above. Thus, per Proposition \\ref{prop:thepoly}, $\\Omega_n(z)$ is the unique monic polynomial with the roots \\eqref{eq:roots}.\n\nReaders familiar with the generalized Fibonacci and Lucas polynomials (\\cite{bergumhoggatt}, \\cite{hoggattbicknell}, \\cite{swamy}, \\cite{webbparberry}) will recognize in \\eqref{eq:lucasrec} the recurrence defining these polynomials, up to the sign change $Y\\mapsto -Y$. The first two polynomials $L_0 = 2$ and $L_1 = X$ coincide with the generalized Lucas polynomials. It is then immediate, since $L_0$ and $L_1$ do not involve $Y$, that\n\n\\begin{prop}\\label{prop:lucas}\nThe $L_n(X,Y)$ of Definition \\ref{def:L} is the image of the $n$th generalized Lucas polynomial $V_n(X,Y)$ under the substitution $Y\\mapsto -Y$.\\qed\n\\end{prop}\n\nThus we can get a formula for $L_n$ (and thus for $\\Omega_n$) from the known formula (\\cite[equation (2.22)]{swamy}) for the generalized Lucas polynomials:\n\\[\nL_n(X,Y) = \\sum_{r=0}^{\\lfloor n/2\\rfloor} (-1)^r\\frac{n}{n-r}\\binom{n-r}{r} X^{n-2r}Y^r\n\\]\nWe include this formula for interest, although it turns out not to be needed for proving the intended result of Price about the product of elliptical chords.\n\nProposition \\ref{prop:lucas} is a reformulation of a standard fact about the generalized Lucas polynomials. The so-called \\textit{Binet formula} (e.g. \\cite[equation (14)]{bergumhoggatt} or \\cite[equation (2.2)]{swamy}) asserts that\n\\begin{equation}\\label{eq:binet}\nV_n(X,Y) = a^n + b^n,\n\\end{equation}\nwhere $a=(X + \\sqrt{X^2 + 4Y})/2$ and $b=(X-\\sqrt{X^2+4Y})/2$, or, equivalently, $X = a+b$ and $Y = -ab$. Normally, one sees $X,Y, V_n$ as conceptually prior to $a,b$. The usual proof is either by a mechanical induction, or via a standard exercise in finding a closed form for the linear recurrence $V_n = XV_{n-1} + YV_{n-2}$. In the latter case, $a,b$ enter as the roots of the characteristic polynomial of that recurrence. However, in our present context, which treats $a,b$ as the conceptual starting point (as in Definition \\ref{def:L}), \\eqref{eq:binet} asserts that up to the substitution $Y\\mapsto -Y$, $V_n(X,Y)$ is the polynomial that expresses the power sum $a^n+b^n$ in terms of the elementary symmetric polynomials $X = a+b$ and $-Y = ab$. This is the content of proposition \\ref{prop:lucas}. Thus our argument for proposition \\ref{prop:lucas} is an alternative proof of the Binet formula. It is essentially computation-free in the sense that the important calculation was subcontracted to Newton's theorem.\n\n\\section{Stretched chords}\\label{sec:solution}\n\nWe apply the above to re-derive Price's formula for the product of the elliptical chords. By Proposition \\ref{prop:thepoly}, the polynomial with roots in the form \\eqref{eq:roots} is $$\\Omega_n(z) = L_n(z,ab) - (a^n+b^n).$$ This polynomial plays the role played by $z^n-1$ in the product of chords in the circle, and $a+b$ is the image of $1$ under the scaling, so by the same logic used in the introduction, the product of the elliptical chords is the absolute value of\n\\[\n\\frac{\\Omega_n(z)}{z-(a+b)}\n\\]\nevaluated at $z=a+b$. This is equal to $\\Omega_n'(a+b)$, since $a+b$ is a root of $\\Omega_n(z)$. We have that\n\\[\n\\Omega_n'(z) = \\frac{d}{dz} L_n(z,ab) = \\frac{d}{dz} V_n(z,-ab)\n\\]\nwhere $V_n$ is the generalized Lucas polynomial, with the second equality by Proposition \\ref{prop:lucas}. Let $U_n(X,Y)$ denote the generalized Fibonacci polynomial, which is defined by the same recurrence as the Lucas polynomials but with the initial data $U_0 = 0$, $U_1= 1$. By a known relation \\cite[equation (3.10)]{swamy} between Lucas and Fibonacci polynomials,\n\\[\n\\frac{d}{dz} V_n(z,-ab) = n U_n(z,-ab)\n\\]\nwhere $U_n$ is the generalized Fibonacci polynomial. Substituting $a+b$ for $z$, the Binet formula for generalized Fibonacci polynomials (\\cite[Theorem 1]{webbparberry} or \\cite[equation (2)]{hoggattbicknell\n) states that\n\\[\nU_n(a+b,-ab) = \\frac{a^n-b^n}{a-b}.\n\\]\nRecalling that $a>|b|\\geq 0$, this is positive, so taking absolute values has no effect. Thus, we recover Price's beautiful result that \n\\begin{prop}[Price \\cite{price1}]\\label{prop:price}\nThe product of the elliptical chord lengths described in the introduction is\n\\[\n\\pushQED{\\qed}\nn\\frac{a^n-b^n}{a-b}.\\qedhere\n\\popQED\n\\]\n\\end{prop}\nAs noted by Price in \\cite{price2}, this yields $nF_n$, as in Table \\ref{tbl:ellipseproducts}, when $a=\\phi$ and $b=\\hat\\phi$, the golden ratio and its algebraic conjugate.\n\n\\section{Solution of $\\Omega_n(z)$ by radicals}\\label{sec:solnbyradicals}\n\nIn this section we bring the inquiry full circle by returning to Cardano. The key polynomial $\\Omega_n(z)$ of our argument is modeled on Cardano's reduced cubic (cf. section \\ref{sec:cardano}). Thus it is no surprise that it can be solved in radicals by essentially the same method. We recover the known formula for the roots of the Lucas polynomials as a special case.\n\nFor $n=3$, we have\n\\[\n\\Omega_3(z) = z^3 - 3abz - (a^3 +b^3)\n\\]\nby comparing \\eqref{eq:identity} with definition \\ref{def:L} and proposition \\ref{prop:thepoly} (which defines $\\Omega_n(z)$). In section \\ref{sec:cardano}, we identified this with Cardano's reduced cubic $x^3 + px + q$ by setting $-3ab = p$ and $-(a^3+b^3)=q$. The presumption is that $p$ and $q$ belong to a prespecified field, such as the rational numbers. Rationality would not have been affected by instead setting $ab = p$ and $a^3 + b^3=q$, so moving forward, this is the convention we generalize: let $p=ab$ and let $q=a^n+b^n$. Then $\\Omega_n(z)=L_n(z,p)-q$ is a univariate polynomial over the field that contains $p$ and $q$, and the goal of a solution in radicals is an expression for its roots in terms of $p$ and $q$.\n\nWe already know the roots are given by \\eqref{eq:roots}; thus we only need to express $a,b$. The method is identical to that which derived Cardano's formula in section \\ref{sec:cardano}. We know that $a^nb^n=p^n$. Thus, $a^n$ and $b^n$ are roots of the quadratic equation\n\\[\nX^2 - qX + p^n = 0.\n\\]\nApplying the quadratic formula and taking $n$th roots, we obtain\n\\begin{equation}\\label{eq:solnbyradicals}\na,b = \\sqrt[n]{\\frac{q \\pm \\sqrt{q^2-4p^n}}{2}}.\n\\end{equation}\nSo a radical expression for the roots of $\\Omega_n(z) = L_n(z,p)-q$ is obtained by substituting \\eqref{eq:solnbyradicals} in \\eqref{eq:roots}.\n\nThe generalized Lucas polynomials are $V_n(X,Y) = L_n(X,-Y)$, i.e. the case $p=-Y$, $q=0$. In this case, \\eqref{eq:solnbyradicals} simplifies to \n\\[\na,b = \\sqrt[n]{\\pm\\sqrt{-(-Y)^n}} = \\sqrt[n]{\\pm i\\sqrt{-Y}^n} = \\xi\\sqrt{-Y}, \\xi^{-1}\\sqrt{-Y},\n\\]\nwhere $\\xi$ is an $n$th root of $i$, i.e. a $4n$th root of unity that is not a $2n$th root. We can take $\\xi = e^{\\pi i / 2n}$, $\\zeta = \\xi^4$, for example, and then by \\eqref{eq:roots}, the roots of $V_n(X,Y)$ (as a polynomial in $X$) are\n\\[\n\\sqrt{-Y} (\\zeta^j\\xi + \\zeta^{-j}\\xi^{-1})= 2\\sqrt{-Y} \\cos \\frac{(1+4j)\\pi}{2n}\n\\]\nfor $j=0,1,\\dots,n-1$. The expression $(1+4j)\\pi/2n$ ranges over all the $(1\\operatorname{mod} 4)$-multiples of $\\pi / 2n = 2\\pi / 4n$ in the range $[0,2\\pi)$. Exploiting the symmetry $\\cos \\theta = \\cos(2\\pi - \\theta)$, those in the range $[\\pi,2\\pi)$ can be replaced by the $(3\\operatorname{mod} 4)$-multiples in the range $[0,\\pi)$. Thus a slight simplification is\n\\[\nX = 2\\sqrt{-Y}\\cos\\frac{(1+2j)\\pi}{2n}\n\\]\nfor $j=0,1,\\dots,n-1$. Compare \\cite[equation (2.24)]{swamy}. The classical Lucas polynomials are the case $Y=1$, so $\\sqrt{-Y} = i$. Compare \\cite[p.~274]{hoggattbicknell}\n\n\\section{History of the circle problem}\\label{sec:circlehistory}\n\nPrice's inspiration, the circle problem described in the introduction, has its own interesting print history, of which we share some highlights here. We think it is likely that the problem has been rediscovered many times, so we make no attempt to be exhaustive. As given in the introduction, the proof (henceforth, the ``standard proof\") that the product of the chord lengths is $n$ (henceforth, the ``circle theorem\") consists of five steps:\n\\begin{enumerate}[label = (\\alph*)]\n\\item Interpret the $n$ equidistant points on the unit circle as the $n$th roots of unity $1,\\zeta,\\dots,\\zeta^{n-1}$ in the complex plane $\\mathbb{C}$.\\label{step:demoivre}\n\\item Interpret the relevant chord lengths as absolute values of the complex numbers $1-\\zeta^j$, $j=1,\\dots,n-1$.\\label{step:geom}\n\\item Exploit the fact that products commute with absolute values in $\\mathbb{C}$ to identify the product of the chord lengths with $(1-\\zeta)\\dots(1-\\zeta^{n-1})$.\\label{step:absval}\n\\item Note that this is the polynomial $(z-\\zeta)\\dots(z-\\zeta^{n-1})$, evaluated at $z=1$.\\label{step:poly}\n\\item Invoke the identity $\\left(z-\\zeta\\right)\\dots(z-\\zeta^{n-1}) = z^{n-1} + \\dots + z + 1$.\\label{step:identity}\n\\end{enumerate}\nTaken individually, each of these is routine, very classical, or both. Thus a broadly-construed ``history of the circle theorem\" would be a history of the complex plane, the roots of unity, and their connection to the circle. This is beyond our present scope. But we mention some striking signposts.\n\nThe first of these is a result quite close to the circle theorem that was discovered in 1716 by Roger Cotes, and enunciated (without proof) in his {\\em Harmonia Mensurarum}, published posthumously in 1722 (see \\cite[p.~194-195]{stillwell}). Cotes is better known as the editor of the second edition of Newton's {\\em Principia}. The following statement of Cotes' theorem is taken from \\cite[p.~195]{stillwell}:\n\n{\\em If $A_0,\\dots,A_{n-1}$ are equally spaced points on the unit circle with center $O$, and if $P$ is a point on $OA_0$ such that $OP=x$, then\n\\[\nPA_0\\cdot PA_1\\cdot \\dots \\cdot PA_{n-1} = 1 - x^n.\n\\]}\nOne obtains the circle theorem by dividing through by $PA_0 = 1-x$ and then letting  $P\\to A_0$ (equivalently $x\\to 1$). \n\nHow Cotes came to this conclusion has not been preserved, but he and his contemporaries Johann Bernoulli and Abraham de Moivre were taking halting steps toward the formula\n\\begin{equation}\\label{eq:demoivre}\n(\\cos \\theta + i\\sin\\theta)^n = \\cos n\\theta + i\\sin n\\theta.\n\\end{equation}\nThis formula now bears de Moivre's name \\cite[p.~192--195]{stillwell}, although it never appeared in his writings \\cite[p.~193]{stillwell}. To a modern reader, and indeed by later in the 18th century, de Moivre's formula would be seen as a consequence of the formula\n\\[\ne^{i\\theta} = \\cos\\theta + i\\sin \\theta\n\\]\ngiven by Euler \\cite[Ch. 8, article 138]{euler} in 1748.\n\nJohn Stillwell has speculated \\cite[p.~195]{stillwell} that Cotes may have used reasoning related to some variant of \\eqref{eq:demoivre}. To us, this formula is the starting point of the circle theorem, for it gives the identification \\ref{step:demoivre} of the roots of unity with points on a unit circle: set the right side of \\eqref{eq:demoivre} to $1$, obtaining $n\\theta=2\\pi k$ for some integer $k$, and then solve for $\\theta$, yielding $\\theta = 2\\pi k / n$. Then consult the left side to conclude $(\\cos 2\\pi k / n + i\\sin 2\\pi k / n)^n = 1$, so that the $n$th roots of unity are the numbers $\\cos 2\\pi k/n + i\\sin 2\\pi k / n$. By the late 18th century, this was a standard maneuver -- see for example \\cite[article 23, p.~249]{lagrange}.\n\nWe are about to skip ahead 200 years, but we mention two major 19th century developments that are mathematically adjacent to our story. \n\nFirst, the identity \\ref{step:identity} is the starting point for the chapter on cyclotomic (``circle-dividing\") equations in Carl Friedrich Gauss' 1801 {\\em Disquisitiones Arithmeticae} (\\cite[Ch. 7]{gauss}). In this work, Gauss proved that any root of unity $\\zeta$ can be expressed in radicals, via calculations in what we now recognize as the Galois group of the polynomial $\\Pi_n(z)=z^{n-1} + \\dots + z + 1$. This was the first real use of Galois groups -- 30 years before Galois! As a corollary, he deduced that if $n$ is prime, a regular $n$-gon can be constructed with ruler and compass if and only if $n$ has the form $2^k + 1$, and therefore the 17-gon is constructible.\n\nSecond, the special case of \\ref{step:identity} with the substitution $z=1$ (as in \\ref{step:poly}, \\ref{step:absval}) figures in the series of monumental papers by Ernst Kummer (\\cite{kummerw}) that birthed the field of algebraic number theory. Kummer investigated how factorization changes when one expands from the ring of integers $\\mathbb{Z}$ to the bigger ring $\\mathbb{Z}[\\zeta]$ consisting of integer linear combinations of $\\zeta$ and its powers. (See \\cite[Ch. 4-6]{edwards} for a modern exposition of Kummer's contribution.) When $n=p$ is prime, substituting $z=1$ into the identity \\ref{step:identity} yields $p=(1-\\zeta)\\dots(1-\\zeta^{p-1})$, which is the factorization of $p$ into primes in $\\mathbb{Z}[\\zeta]$ (e.g. \\cite[p.~174]{kummer}).\n\nWe now reach the 20th century. The circle theorem was stated in a short 1954 note of W. Sichardt \\cite{sichardt} in \\textit{Zeitschrift f\\\"{u}r Angewandte Mathematik und Mechanik} (\\textit{ZAMM}). This is the first instance of which we are aware in which the theorem was given in the form we have stated it, although again we make no claim to comprehensivity.\n\nSichardt made the observation much earlier --- in 1927 --- in the context of an engineering problem having to do with the construction of wells! The calculation of this particular product of chords was directly motivated by the engineering context.\n\nSichardt described finding the pattern in the products empirically. He did not claim credit for the proof. He stated that an unnamed mathematician who worked for Siemens provided a proof that was lost during the war. He attributed the proof given in the note to a Prof. Szab\\'{o} of the Technical University of Berlin. It is a slightly less clean version of the standard proof. It makes use of the fact that the length of a chord in the unit circle subtended by an angle $\\theta$ is $2\\sin(\\theta/2)$, and thus the product of interest can be expressed as\n\\[\n2^{n-1}\\prod_{j=1}^{n-1} \\sin \\pi j/n.\n\\]\nEach factor $\\sin \\pi j / n$ is expressed as $\\frac{1}{2i}\\left(\\eta^j - \\eta^{-j}\\right)$, where $\\eta$ is a $2n$th root of unity rather than an $n$th, so a certain amount of bookkeeping is needed to arrive at the expression in terms of the polynomial $\\Pi_n(z)$. The comparative cleanness of the standard proof comes from the fact that taking absolute values allows one to forego all this bookkeeping.\n\nIn 1972, Kurt Eisemann, in another short note in {\\em ZAMM} \\cite{eisemann}, extended Sichardt's result by considering the product of the lengths of perpendiculars from the center of the circle to the chords. Independently, Zalman Usiskin \\cite{usiskin} in 1979 derived various identities involving products of sines using only the standard trigonometric identities $\\sin 2\\theta = 2\\sin\\theta\\cos\\theta$ and $\\cos\\theta = \\sin (\\pi/2-\\theta)$, and noted the interpretation of these identities in terms of chord lengths. One of the identities proven by Usiskin yielded the case $n=45$ of the circle theorem, and he stated the general case without proof.\n\nIn 1987, Steven Galovich \\cite{galovich} took up Usiskin's challenge to find a general principle explaining his various sine-product identities. Among other results, Galovich proved the circle theorem, using exactly the standard proof. But he introduced it with the words, ``Although the next theorem and proof are evidently well known, it is natural to include them in this note.\" \\cite[p.~112]{galovich} \n\nBy this point, the circle theorem had been posed as an exercise in textbooks, for example \\cite[p.~16, exercise 12]{baknewman} and \\cite[p.~69, problem 44]{kreyszig}.\n\nIn 1995, Andre Mazzoleni and Samuel Shan-Pu Shen published a note \\cite{mazzolenishen} in the February issue of {\\em Mathematics Magazine} giving a short proof of the circle theorem via the theory of residues of a function of a complex variable. For precedents for the result, they cited exercises in several textbooks in complex analysis. Responses from readers were published in the June and October issues pointing out other precedents. Among these readers were Usiskin, and also Eisemann, who called attention to Sichardt's contribution as well as his own. One reader fit an extremely concise version of the standard proof into a letter to the editor \\cite{silva}.\n\nIn 2002, Barry Lewis gave a riff \\cite{lewis} on the circle theorem by deriving formulas for power sums of the chord lengths, rather than their product.\n\nFinally we come to Price's work in the early 2000's. In addition to \\cite{price1}, \\cite{price2} which are the starting point of the present article, Price also produced \\cite{price3}, extending Eisemann's \\cite{eisemann} results to the elliptical situation. In these works, Price cited Sichardt for the circle theorem, and gave the standard proof.\n\n\\section{Further remarks}\\label{sec:furtherremarks}\n\n\\subsection{The Binet formula for the generalized Fibonacci polynomials}\\label{sec:binetfibonacci}\n\nWe touted our computation-free proof of the Binet formula for the generalized Lucas polynomials in section \\ref{sec:newtonbinet}, but then in section \\ref{sec:solution}, we {\\em used} the Binet formula for the generalized Fibonacci polynomials, cited from the literature and so justified by one of the standard proofs, in our proof of Price's main result (Proposition \\ref{prop:price}). This situation calls out for an independent proof of the Binet formula for generalized Fibonacci polynomials along the lines of what we have done above for Lucas polynomials. Can this be given?\n\nIt can. The proof above for the Lucas polynomials proceeds by defining (Definition \\ref{def:L}) a polynomial that expresses $a^n+b^n$ in terms of $a+b$ and $ab$, whose existence and uniqueness are guaranteed by the fundamental theorem on symmetric polynomials; verifying that it is the generalized Lucas polynomial in the cases $n=0$ and $n=1$; and then specializing Newton's theorem to prove that it obeys the Fibonacci/Lucas recursion (up to a sign change) and therefore (Proposition \\ref{prop:lucas}) coincides with the generalized Lucas polynomial (up to the same sign change) for all $n$. Finally it observes that the Binet formula can be interpreted as the statement that the generalized Lucas polynomial expresses $a^n+b^n$ in terms of $a+b$ and $-ab$, i.e. the statement that has just been proven. In the Fibonacci case, the Binet formula states that\n\\[\nU_n(X,Y) = \\frac{a^n-b^n}{a-b},\n\\]\nwhere $a,b$ are defined by $a+b=X$, $ab=-Y$, as above. Now $(a^n-b^n)/(a-b)$ is equal to $$a^{n-1}+a^{n-2}b+\\dots+ab^{n-2} + b^{n-1},$$ the {\\em complete homogeneous symmetric polynomial} of degree $n-1$ in $a,b$. Like the power sum $a^n+b^n$, the fundamental theorem on symmetric polynomials guarantees a unique polynomial expressing this in terms of $a+b$ and $ab$, so we can copy the entire pattern of the above proof if we can find an analogue to Newton's theorem that tells us this polynomial (call it $F_n(X,Y)$) obeys the Fibonacci/Lucas recursion up to the sign change. (The verification that it coincides with $U_n$ if $n=0$ and $n=1$ is trivial: $(a^0-b^0)/(a-b) = 0$, so $F_0 = 0= U_0$, and $n=1$ is similar.) \n\nIn fact, there is such an analogue. For any $m$ indeterminates $a_1,\\dots,a_m$, and any natural number $j$, let $h_j$ be the complete homogeneous symmetric polynomial of degree $j$, i.e. the sum of {\\em every} monomial in the $a_i$'s of degree $j$. Let $\\sigma_j$ be the elementary symmetric polynomial of degree $j$, as in section \\ref{sec:newtonbinet}. Then these polynomials satisfy a relation nearly identical to the one between the $\\sigma$'s and the power sums captured by Newton, namely\n\\begin{equation}\\label{eq:completehom}\nh_n-h_{n-1}\\sigma_1+h_{n-2}\\sigma_2 - \\dots \\pm h\\sigma_{n-1} \\mp \\sigma_n = 0\n\\end{equation}\nfor all $n\\geq 1$. (As in section \\ref{sec:newtonbinet}, this continues to hold for $n>m$ via the convention that $\\sigma_j=0$ for $j>m$.) In the two-variable case, where we have $(a^k-b^k)/(a-b) = h_{k-1}$, this specializes to\n\\[\n\\frac{a^n-b^n}{a-b} - (a+b) \\frac{a^{n-1}-b^{n-1}}{a-b} + (ab)\\frac{a^{n-2}-b^{n-2}}{a-b} = 0\n\\]\nfor all $n\\geq 3$. In fact it even specializes for $n = 2$ in view of $(a^0-b^0)/(a-b)=0$ so that the final term vanishes. The proof now proceeds exactly as in section \\ref{sec:newtonbinet} to identify $F_n(X,Y)$ with $U_n(X,-Y)$.\n\nThe formula \\eqref{eq:completehom} is easy to find in the literature (e.g. \\cite[equation (7.13)]{stanley}), but for the sake of self-containedness and because it is very beautiful, we give the usual proof via formal power series. Let\n\\[\nH(t) = \\sum_0^\\infty h_jt^j\n\\]\nbe the ordinary generating function for the complete homogeneous symmetric polynomials. Because $h_j$ is the sum of all monomials of a given degree $j$ in the given set of indeterminates, $H(t)$ is the sum of {\\em all} monomials, each multiplied by $t^{\\text{its degree}}$. Thus\n\\begin{align*}\nH(t) &= \\left(1 + a_1t + a_1^2t^2 + \\dots\\right) \\ldots \\left(1 + a_mt + a_m^2t^2 + \\dots\\right)\\\\\n&= \\left(\\frac{1}{1-a_1t}\\right)\\ldots\\left(\\frac{1}{1-a_mt}\\right).\n\\end{align*}\nThen\n\\[\nH(t)^{-1} = \\left(1 - a_1t\\right) \\ldots \\left(1 - a_mt\\right) = \\sum_0^\\infty (-1)^k\\sigma_k t^k,\n\\]\nsince the coefficient of a given $t^k$ is the sum of all products of $k$-sets of $-a_j$'s, by multiplying out. Thus,\n\\[\n1 = H(t)H(t)^{-1} = \\sum_0^\\infty h_jt^j\\sum_0^\\infty (-1)^k\\sigma_k t^k,\n\\]\nand \\eqref{eq:completehom} is obtained from this formula by extracting the coefficient of $t^n$ on both sides and comparing.\n\n\\subsection{Relation between generalized Fibonacci and Lucas polynomials}\\label{sec:fiblucas}\n\nIn the proof of Price's formula, Proposition \\ref{prop:price}, we made use of the following identity relating generalized Fibonacci and Lucas polynomials, for which we cited a paper \\cite{swamy} by Swamy:\n\\[\n\\frac{d}{dX} V_n(X,Y) = nU_n(X,Y).\n\\]\nIn the interest of keeping this paper self-contained, we give a proof. It resembles a calculation in \\cite[p.~152]{price2}. (It is different from the proof given by Swamy, which is based on generating functions.) Set $a+b=X$ and $ab=-Y$ as usual. Treating $Y$ as a constant, $b$ and therefore $X$ and $V_n(X,Y)$ become rational functions of $a$. We have $b=-Ya^{-1}$, so that $db/da = Ya^{-2} = -ba^{-1}$. Thus, by the Binet formulas for $V_n$ and $U_n$ and the chain rule, we have\n\\begin{align*}\n\\frac{d}{dX} V_n(X,Y) &= \\frac{dV_n(X,Y)/da}{dX/da}\\\\\n&= \\frac{d(a^n + b^n)/da}{d(a+b)/da}\\\\\n&= \\frac{na^{n-1} +nb^{n-1}(-ba^{-1})}{1-ba^{-1}}\\\\\n&= n\\frac{a^n - b^n}{a-b}\\\\\n&= nU_n(X,Y).\n\\end{align*}\n\n\\subsection{Relation with Price's work}\\label{sec:price}\n\n\nA comparison of the method above with \\cite{price1} and \\cite{price2} has a through-the-looking-glass quality. There are many contact points, but the overall effect is completely different. The direct calculations in \\cite{price1} and \\cite{price2} are here contextualized as special cases of classical theorems and other known results. What follows is our best attempt to tease out the relationship in more detail.\n\nPrice \\cite{price1} worked with a polynomial $P_n(z)$ (called $P_n(z;a,b)$ in \\cite{price2}) that is equal to our $L_n(z,ab)$, but {\\em defined} it to be the polynomial such that $P_n(z) - (a^n+b^n)$ has the roots \\eqref{eq:roots}. Our additions are the direct proof, using the invariance of $ab$ and $a^n+b^n$ under $a\\mapsto \\zeta a, b\\mapsto \\zeta^{-1}b$, that the polynomial characterized by Definition \\ref{def:L} is this same polynomial, and the related observation that it generalizes Cardano's reduced cubic.\n\nPrice proved the recurrence relation $P_n(z) = zP_{n-1}(z) - ab P_{n-2}(z)$ in \\cite{price1}. This is our \\eqref{eq:lucasrec}, after the substitution $Y\\mapsto ab$. Price's proof amounts to verifying a version of \\eqref{eq:newton2} with direct calculation. In \\cite{price2}, he observed that this makes $P_n(z)$ a generalized Lucas polynomial, and used this to prove the Binet formulas for Lucas and Fibonacci numbers. Our additions are the observation that \\eqref{eq:newton2} is immediate from Newton's theorem, that the analogous formula for Fibonacci polynomials is immediate from a standard fact about complete homogeneous symmetric polynomials (section \\ref{sec:binetfibonacci}), and the interpretation of the Binet formulas as the statements that the Lucas and Fibonacci polynomials are the polynomials that express, respectively, power sums and complete homogeneous symmetric polynomials, in terms of elementary symmetric polynomials.\n\nPrice's \\cite{price1} proof of proposition \\ref{prop:price} proceeds by first proving with an induction based on the same calculation mentioned in the previous paragraph that, for any $\\theta\\in [0,2\\pi)$,\n\\begin{equation}\\label{eq:continuous}\nP_n(ae^{i\\theta} + b e^{-i\\theta}) = a^ne^{in\\theta} + b^ne^{-in\\theta},\n\\end{equation} and then deriving proposition \\ref{prop:price} from a computation with the continuous variable $\\theta$ involving the L'Hopital rule. The L'Hopital calculation is repeated in \\cite[p.~152]{price2}. We do not need \\eqref{eq:continuous} for our proof of \\ref{prop:price}, but it is immediate from our work by substituting $a\\mapsto ae^{i\\theta}$ and $b\\mapsto be^{-i\\theta}$ in $L_n(a+b,ab) = a^n + b^n\n. We instead derived proposition \\ref{prop:price} from the Binet formulas and the known identity $\\frac{d}{dX}V_n(X,Y) = nU_n(X,Y)$. In the name of keeping this paper self-contained, we included a proof of this identity, which is structurally similar to Price's L'Hopital calculation.\n\nAnother addition is our observation (section \\ref{sec:solnbyradicals}) that Cardano's method allows us to express the roots of $\\Omega_n(z)$ by radicals and thereby obtain the known formula for the roots of the Lucas polynomial.\n\nProposition \\ref{prop:price}, and the other results given here, do not exhaust the results found in \\cite{price1} and \\cite{price2}. In \\cite{price1}, Price also considered the products of chord lengths that arise from rotating the roots of unity by a fixed angle along the unit circle prior to scaling. In \\cite{price2}, he used the interpretation of Fibonacci and Lucas numbers in terms of products of elliptical chord lengths to recover identities and divisibility properties of these numbers, such as $F_{2n} = F_nL_n$.\n\n\\section{Acknowledgement}\n\nSeveral individuals were involved in calling the authors' attention to Price's work: Francis Su, who included the special case of Table \\ref{tbl:ellipseproducts} in his Harvey Mudd Math Fun Facts (\\cite{su}); Bowen Kerins and Darryl Yong, who then included the problem of computing this product in a problem set during a summer course at the Park City Mathematics Institute; and Sam Shah, who posted the problem on his blog, whereby the authors were acquainted with it. The authors wish to thank Kerins in particular, who was a generous correspondent, and also Thomas Price himself, Tom Edgar, and Harold Edwards, for useful comments and for alerting us to \\cite{price3}, \\cite{galovich}, and \\cite{kummerw} respectively.\n\n", "meta": {"timestamp": "2019-09-25T02:04:06", "yymm": "1810", "arxiv_id": "1810.00492", "language": "en", "url": "https://arxiv.org/abs/1810.00492", "abstract": "A beautiful theorem of Thomas Price links the Fibonacci numbers and the Lucas polynomials to the plane geometry of an ellipse, generalizing a classic problem about circles. We give a brief history of the circle problem, an account of Price's ellipse proof, and a reorganized proof, with some new ideas, designed to situate the result within a dense web of connections to classical mathematics. It is inspired by Cardano's solution of the cubic equation and Newton's theorem on power sums, and yields an interpretation of generalized Lucas polynomials in terms of the theory of symmetric polynomials. We also develop additional connections that surface along the way; e.g., we give a parallel interpretation of generalized Fibonacci polynomials, and we show that Cardano's method can be used write down the roots of the Lucas polynomials.", "subjects": "History and Overview (math.HO); Number Theory (math.NT)", "title": "Chords of an ellipse, Lucas polynomials, and cubic equations", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.98866824713641, "lm_q2_score": 0.8577681049901037, "lm_q1q2_score": 0.8480480888100859}}
{"url": "https://arxiv.org/abs/2202.11420", "title": "Sub-optimality of Gauss--Hermite quadrature and optimality of the trapezoidal rule for functions with finite smoothness", "abstract": "The sub-optimality of Gauss--Hermite quadrature and the optimality of the trapezoidal rule are proved in the weighted Sobolev spaces of square integrable functions of order $\\alpha$, where the optimality is in the sense of worst-case error. For Gauss--Hermite quadrature, we obtain matching lower and upper bounds, which turn out to be merely of the order $n^{-\\alpha/2}$ with $n$ function evaluations, although the optimal rate for the best possible linear quadrature is known to be $n^{-\\alpha}$. Our proof of the lower bound exploits the structure of the Gauss--Hermite nodes; the bound is independent of the quadrature weights, and changing the Gauss--Hermite weights cannot improve the rate $n^{-\\alpha/2}$. In contrast, we show that a suitably truncated trapezoidal rule achieves the optimal rate up to a logarithmic factor.", "text": "\\section{Introduction} \\label{sec:intro}\nThis paper is concerned with a sub-optimality of Gauss--Hermite quadrature and an optimality of the trapezoidal rule.\n\nGiven a function $f\\colon \\mathbb{R}\\to \\mathbb{R}$, Gauss--Hermite quadrature is\none of the standard numerical integration methods to compute the integral\n\\begin{equation}\n    I(f):=\\int_\\mathbb{R}f(x)\\frac{1}{\\sqrt{2\\pi}}\\mathrm{e}^{-x^2/2}\\mathrm{d}x.\n    \\label{eq:def_I(f)}\n\\end{equation}\nIt is a Gauss-type quadrature formula, i.e., the quadrature points are the zeros of the degree $n$ orthogonal polynomial associated with the weight function, and the corresponding quadrature weights are readily defined. With the weight function $\\rho(x):=\\frac{1}{\\sqrt{2\\pi}}\\mathrm{e}^{-x^2/2}$, the orthogonal polynomial we have is the (so-called probabilist's) Hermite polynomial.\n\nGauss--Hermite quadrature is widely used; \nhere, we just mention \nspectral methods \\cite{Guo.BY_1999_Hermite,mao2017hermite,Canuto.C_etal_2006_book}, and applications in \naerospace engineering \\cite{braun2021satellite,BBM2020book}, \nfinance \\cite{Brandimarte2006book,FR2008book,GH2010}, and \nphysics \\cite{SS2016book,Gezerlis2020book}.\nNevertheless, the limitation of this method seems to be less known.\n\nWe start with numerical results that illustrate this deficiency. \nFigure~\\ref{fig:GH-sub} shows a comparison of the Gauss--Hermite rule and a suitably truncated trapezoidal rule on $\\mathbb{R}$.  \nHere the target function in \\eqref{eq:def_I(f)} is $f(x)=|x|^p$ with $p\\in \\{1,3,5\\}$. \nFor the trapezoidal rule, we integrate $f(x)\\rho(x)$ with a suitable cut-off of the domain. \nWe discuss the setting of this experiment in more detail at the end of Section~\\ref{sec:trapez}.\n\\begin{figure}[H]\n\\centering\n \\includegraphics[scale=1.05]{graph-fitted-tri.pdf}\n \\vspace*{-7mm}\n\\caption{Absolute integration errors for $f(x)=|x|^p$ with $p=1$ \\textup{(}left{)}, $p=3$ \\textup{(}centre{)}, and $p=5$ \\textup{(}right{)}.\nGauss--Hermite rule exhibits a slower error decay than the trapezoidal rule does.}\n \\label{fig:GH-sub}\n\\end{figure}\n\nWhat we observe in Figure~\\ref{fig:GH-sub} is that while the trapezoidal rule achieves around $\\mathcal{O}(n^{-p-0.8})$, Gauss--Hermite quadrature achieves only a slower convergence rate almost $\\mathcal{O}(n^{-p/2-0.5})$, where $n$ is the number of quadrature points. \nA similar empirical inefficiency of Gauss--Hermite quadrature is reported in a paper by one of the present authors and Nuyens~\\cite{NS2021},  also in a recent paper by Trefethen~\\cite{T2021}\nargued that Gauss--Hermite quadrature converges more slowly than the truncated trapezoidal\nrule as $n\\to\\infty$, because its quadrature points are unnecessarily spread out, and in effect not enough quadrature points are utilised.\n\nIn this paper, we \n\\emph{prove} a sub-optimality of Gauss--Hermite quadrature. \nMore precisely, we establish a sharp lower bound for the error decay of Gauss--Hermite quadrature in the sense of worst case error. Moreover, we show that a suitably truncated trapezoidal rule achieves the optimal rate of convergence, up to a logarithmic factor.\n\\iffalse\n\\yk{For analytic functions, efficiency of the trapezoidal rule is well known;  \nrelated studies date back at least to the paper by Goodwin in 1949 \\cite{Goodwin1949}, and this accuracy is not only widely known to, but also still actively studied by contemporary numerical analysts  \\cite{S1997,MS2001,GilEtAl2007,Waldvogel2011,TW2014,T2021}.\nSugihara \\cite{S1997} conducted an extensive research on the integration error for the trapezoidal rule corresponding to various  decay conditions. ...in particular for functions decaying like $\\exp(-x^2)$ he shows the rate $\\exp(-Cn...)$..the constant is explicit. \n\\ys{(How about here...?)Our results show that this efficiency extends to Sobolev class functions, where we do not have tools from complex analysis such as contour integrals. \nOur proof uses the strategy recently developed by one of the present authors and Nuyens~\\cite{NS2021} for a class of quasi-Monte Carlo methods.}\nThere are also lower bounds for a general class quadratures. These are not matching lower bounds but close, showing the near optimality of the trapezoidal rule in this function space.\n\\cite{T2021} shows .....\nFor GH, Barret (REF) seems to be the first to have considered the rate...\nalso Davis and ....book (4.6.1.18)\nNote that there is no decaying condition written anywhere. We are not aware of any.\nTrethenn SIREV's  numerical result (Figure something) support this rate for functions decaying like $\\exp(-x^2)$. \nIs TR optimal?\nHow do TR and GH compare for analytic functions? We do not address these important questions in this paper.\n} \n\\fi\n\nThe integrands of our interest are functions with finite smoothness. \nIn this regard, we will work under the assumption that the function $f$ lives in the weighted $L^2$-Sobolev space with smoothness $\\alpha\\in\\mathbb{N}$.  \nThis function space is widely used; see for example the books \\cite{B1998,Canuto.C_etal_2006_book,Shen.J_2011_book} and references therein. \nFor $\\alpha\\in\\mathbb{N}$, this space is equivalent to the so-called Hermite space of finite smoothness, which has been attracting attention in high-dimensional computations; see \\cite{IG2015_Hermite,DILP2018,gnewuch2021countable} for its use in high-dimensional computations, and see \\cite[1.5.4. Proposition]{B1998} together with Section~\\ref{sec:space} below for the equivalence.\n\nIn this setting, we prove that the worst case integration error of Gauss--Hermite quadrature is bounded from below by  $n^{-\\alpha/2}$ up to a constant. This rate matches the upper bound shown by Mastroianni and Monegato \\cite{MM1994}, and thus cannot be improved. \nMoreover, this rate provides a rigorous verification of the numerical findings  \\cite[Section 4]{DILP2018}, where the authors computed  approximate values of the worst-case error for Gauss--Hermite quadrature in the Hermite space of finite smoothness, and observed the rate $\\mathcal{O}(n^{-\\alpha/2})$ for  $\\alpha=1,2,$ and $3$.\n\nIn the proof, we exploit the structure of the Gauss--Hermit quadrature \\textit{points}. The argument is independent of the quadrature weights, and thus tuning them does not change the result. \nThe proof in particular indicates that, if the spacing of a node set decreases asymptotically no faster than $1/\\sqrt{n}$, then the corresponding quadrature rule cannot achieve the worst-case error better than $O(n^{-\\alpha/2})$; see the proof of Lemma~\\ref{lem:lower_bound_general} together with  Theorem~\\ref{thm:GH-LB}.\n\nIt turns out that this rate is merely half of the best possible: \nif we allow $n$ quadrature-points and weights to be arbitrary, then the best achievable using (linear) quadrature is of the rate $\\mathcal{O}(n^{-\\alpha})$; see \\cite[Theorem 1]{DILP2018} for a precise statement.\nDick et al.~\\cite{DILP2018} also show that a class of numerical integration methods based on so-called (scaled) higher order digital nets achieve the optimal rate $\\mathcal{O}(n^{-\\alpha})$ up to a logarithmic factor in the multi-dimensional setting, including the one-dimensional setting as a special case.\n\nOur results on the trapezoidal rule show that the trapezoidal rule, a method arguably significantly simpler than one-dimensional higher order digital nets,\nalso achieves this optimal rate, up to a logarithmic factor, and thus is nearly \\textit{twice} as fast as the error-decay rate of Gauss--Hermite rule. \nIt is also worth mentioning that Gauss--Hermite quadrature requires a nontrivial algorithm to generate quadrature points, whereas for the trapezoidal rule we simply have equispaced points.\n\nFor analytic functions, the efficiency of the trapezoidal rule is well known;  \nrelated studies date back at least to the paper by Goodwin in 1949 \\cite{Goodwin1949}, and this accuracy is not only widely known, but also still actively studied by contemporary numerical analysts \\cite{S1997,MS2001,GilEtAl2007,Waldvogel2011,TW2014,T2021}. \nOur results show that this efficiency extends to Sobolev class functions, where we do not have tools from complex analysis such as contour integrals. \nOur proof uses the strategy recently developed by one of the present authors and Nuyens~\\cite{NS2021} for a class of quasi-Monte Carlo methods.\n\nWe now mention other error estimates for Gauss--Hermite quadrature in the literature.  \nBased on results by Freud~\\cite{Freud1972}, Smith, Sloan, and Opie~\\cite{SSO1983} showed an upper bound $\\mathcal{O}(n^{-\\alpha/2})$ for $\\alpha$-times continuously differentiable functions whose $\\alpha$-th derivative satisfies a suitable growth condition for $\\alpha\\in\\mathbb{N}$. \nSince the weighted Sobolev space seems to be more frequently used, our focus is on this class. \nDella Vecchia and Mastroianni~\\cite{DM2003} showed an upper bound $\\mathcal{O}(n^{-1/6})$ for $\\rho$-integrable a.e.~differentiable functions whose derivative is also $\\rho$-integrable. \nMoreover, their result implies a matching lower bound in the sense of worst-case error, and thus  in this sense their upper bound is sharp. \nIt does not seem to be trivial to determine whether their bounds generalise, for example to the order $n^{-\\alpha/6}$ or to $n^{-\\alpha/2+1/3}$, with the $\\alpha$-times differentiability of the integrand for $\\alpha\\ge2$.\nIn contrast, our results show that, by assuming the square $\\rho$-integrability, the rate improves to $\\mathcal{O}(n^{-\\alpha/2})$, for general $\\alpha\\in\\mathbb{N}$. \n\n\nFor analytic functions, \nthe rate $\\exp(-C\\sqrt{n})$ with a constant $C>0$ has been mentioned  in the literature for functions with a \\textit{suitable decay} \nthat are analytic in a strip region.   \nBarrett~\\cite{Barrett.W_1961} seems to be the first to have presented this rate; see also Davis and Rabinowitz \\cite[Equation~(4.6.1.18)]{DR1984}. \nNote, however, that no explicit proof or statement is given in either of these references; in particular, the decay condition for which this rate holds is not given. \nWe are not aware of any reference that gives an explicit statement with complete assumptions. \nOn the other hand, for the trapezoidal rule, Sugihara~\\cite{S1997} conducted an extensive research on the integration error for functions analytic in a strip region with various decay conditions. \nIn particular, for functions decaying at the rate $\\exp(-\\tilde{C}|x|^\\rho)$ ($\\rho\\geq 1$) on the real axis, under other suitable assumptions he established the rate $\\exp(-{C}^*n^{\\rho/(\\rho+1)})$, with an explicit constant ${C}^*>0$; see \\cite[Theorem 3.1]{S1997} for a precise statement.  \nHence, whatever the decay condition the Gauss--Hermite rule requires to give the rate $\\exp(-C\\sqrt{n})$ may be, for $\\rho\\geq 1$ the trapezoidal rule attains a faster rate for the function class considered in \\cite{S1997}. \nNote that lower bounds for the integration error are also presented in \\cite{S1997}, which shows that the trapezoidal rule is near-optimal in the function class considered there. \nTrefethen \\cite{T2021} makes the comparison of the trapezoidal rule and Gauss--Hermite quadrature explicit. \nIn \\cite[Theorem 5.1]{T2021}, for the trapezoidal rule and various other quadrature rules he established the rate \n$\\exp(-C{n}^{2/3})$ for a class of functions that are analytic in a strip region which decays at the rate $\\exp(-x^2)$ on the real axis. \nHe also presents a numerical result for integrating $\\cos(x^3)\\exp(-x^2)$ with the physicists' Gauss--Hermite rule, which supports the rate $\\exp(-C\\sqrt{n})$ for the Gauss--Hermite rule.\n\nBefore moving on, we mention that our results motivate further studies of Gauss--Hermite based algorithms in high-dimensional problems. \nIntegration problems in high dimensions arise, for example in computing statistics of solutions of partial differential equations parametrised by random variables. \nIn particular, integration with respect to the Gaussian measure has been attracting increasing attention; see for example \\cite{Graham.I_eta_2015_Numerische,C2018,YK2019,HS2019ML,D2021,dung2022analyticity}. \nThe measure being Gaussian, algorithms that use Gauss--Hermite quadrature as a building block have gained popularity \\cite{C2018,D2021,dung2022analyticity}.  \nThe key to proving error estimates in this context is the regularity of the quantity of interest with respect to the parameter, and for elliptic and parabolic problems such smoothness, even an analytic regularity, has been shown  \\cite{BNT2010SIREV,NT2009parabolic,dung2022analyticity}. \nIn contrast, the solutions of parametric hyperbolic systems suffer from limited regularity under mild assumptions \\cite{MNT2013_StochasticCollocationMethod,MNT2015_AnalysisComputationElasticWave}. \nHence, our results, which show the sub-optimality of Gauss--Hermite rule for functions with finite smoothness, caution us and encourage further studies of the use of algorithms based on Gauss--Hermite rule for this class of problems.\n\nFinally, we note that if we have the weight function \n$\\mathrm{e}^{-x^2}$ instead of $\\mathrm{e}^{-x^2/2}$,\nthe corresponding orthogonal polynomials are called physicist's Hermite polynomials. Our results for Gauss--Hermite quadrature can be obtained for these polynomials by simply rescaling our results by $x\\mapsto \\sqrt{2}x.$\nLikewise, results for physicist's Hermite polynomials in the literature, e.g.~in \\cite{Szego1975}, are used throughout this paper. \n\nThe rest of this paper is organized as follows. In Section~\\ref{sec:space} we introduce necessary definitions such as Hermite polynomials, the weighted Sobolev spaces, and the Hermite spaces. \nWe also discuss the norm equivalence between the weighted Sobolev space and the Hermite space. \nIn Section~\\ref{sec:GH-bounds} the sub-optimality of Gauss--Hermite quadrature is shown. In particular, we obtain matching lower and upper bounds for the worst-case error.\nIn Section~\\ref{sec:trapez} the optimality of the trapezoidal rule is shown.\nSection~\\ref{sec:conc} concludes this paper.\n\n\\section{Function spaces with finite smoothness} \\label{sec:space}\nThroughout this paper, we use the weighted space $L^2_\\rho:=L^2_\\rho(\\mathbb{R})$, the normed space consisting of the equivalence classes of Lebesgue measurable functions $f\\colon\\mathbb{R}\\to\\mathbb{R}$ satisfying $\\|f\\|_{L^2_\\rho}^2:=\\int_{\\mathbb{R}}|f(x)|^2\\rho(x)\\mathrm{d}x<\\infty$, where the equivalence relation is given by $f\\sim g$ if and only if $\\|f-g\\|_{L^2_\\rho}=0$.\n\\subsection{Hermite polynomials}\nFor $k\\in\\mathbb{N}\\cup\\{0\\}$, the $k$-th degree probabilist's Hermite polynomial is given by\n\\begin{align}\\label{eq:Hermite-def}\n    H_{k}(x)\n\t=\\frac{(-1)^{k}}\t{\\sqrt{k!}}\n    \t\\mathrm{e}^{x^{2}/2}\n        \\frac{\\mathrm{d}^{k}}{\\mathrm{d} x^{k}}\n        \t\\mathrm{e}^{-x^{2}/2},\\quad x\\in\\mathbb{R},\n\\end{align}\nwhere they are normalised so that $\\|H_k\\|_{L^2_\\rho}=1$ for all $k\\in\\mathbb{N}\\cup\\{0\\}$. \nThe polynomials $(H_k)_{k\\geq 0}$ form a complete orthonormal system for $L^2_\\rho$.\n\nThe following properties are used throughout the paper.\n\\begin{align}\\label{eq:Hermite-deriv}\nH_k'(x)=\\sqrt{k}H_{k-1}(x),\\; k\\ge1;\n\\end{align}\n\\begin{align}\\label{eq:Hermite-rho-deriv}\\\n\\frac{\\mathrm{d}^{\\tau}}{\\mathrm{d} x^{\\tau}}\\left(H_k(x)\\rho(x)\\right)\n&=        \t\n\\frac{(-1)^{k}}\t{\\sqrt{2\\pi k!}}\n    \t\\mathrm{e}^{x^{2}/2}\n        \\frac{\\mathrm{d}^{k+\\tau}}{\\mathrm{d} x^{k+\\tau}}\n        \t\\mathrm{e}^{-x^{2}/2} \\nonumber\n        \t\\\\\n        \t&=\n        \t(-1)^\\tau\\sqrt{\\frac{(k+\\tau)!}{k!}}H_{k+\\tau}(x)\\rho(x),\\; k\\ge0,\\; \\tau\\ge0.\n\\end{align}\n\\subsection{Weighted Sobolev space}\nThe function space we consider is the Sobolev space of square integrable functions, the integrability condition of which is imposed by the Gaussian measure.\n\\begin{Definition}[Weighted Sobolev space]\\label{def:Sobolev}\nFor $\\alpha\\in\\mathbb{N}$, the weighted Sobolev space $\\mathscr{H}_{\\alpha}$(with the weight function $\\rho$) is the class of all functions $f\\in L^2_\\rho$ such that $f$ has weak derivatives satisfying $f^{(\\tau)}\\in L^2_\\rho$ for $\\tau=1,\\dots,\\alpha$:\n\\begin{align*}\n \\mathscr{H}_{\\alpha}:=\n    \\Biggl\\{\n    f \\in L^2_\\rho \\;\\bigg|\\; \\|f\\|_\\alpha := \\biggl(\\sum_{\\tau=0}^\\alpha \\|f^{(\\tau)}\\|^2_{L^2_\\rho}\\biggr)^{1/2} < \\infty\n    \\Biggr\\}.\n\\end{align*}\n\\end{Definition}\nElements in $\\mathscr{H}_{\\alpha}$ for $\\alpha\\in \\mathbb{N}$ are in the standard local Sobolev space $W^{1,2}_{\\mathrm{loc}}(\\mathbb{R})$, and thus admit a continuous representative. \nIn what follows, we always take the continuous representative of $f\\in \\mathscr{H}_{\\alpha}$.\n\nWe recall another important class of functions, the so-called Hermite space.\nFor this space we follow the definition of \\cite{DILP2018}.\n\\begin{Definition}[Hermite space with finite smoothness]\\label{def:Hermite}\nFor $\\alpha\\in\\mathbb{N}$, the Hermite space with finite smoothness $\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ is given by\n\\begin{align*}\n \\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}:=\n    \\Bigg\\{\n    f \\in L^2_\\rho \\;\\bigg|\\; \\|f\\|_{\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}} := \n        \\biggl(\n            \\sum_{k=0}^\\infty r_{\\alpha}(k)^{-1}\\bigl|\\widehat{f}(k)\\bigr|^2\n        \\biggr)^{1/2} < \\infty\n    \\Bigg\\},\n\\end{align*}\nwhere $\\widehat{f}(k)=(f,H_k)_{L^2_{\\rho}}:=\\int_\\bbR f(x)H_k(x) \\rho(x)\\mathrm{d} x$, and\n\\[\nr_\\alpha(k):=\\begin{cases}1, &\\mathrm{if }\\; k=0,\\\\ \\bigl(\\sum^\\alpha_{\\tau=0}\\beta_\\tau(k)\\bigr)^{-1}, &\\mathrm{if }\\;k\\ge1, \\end{cases} \n\\;\\mathrm{and}\\;\n\\beta_\\tau(k):=\\begin{cases} \\frac{k!}{(k-\\tau)!}, &\\mathrm{if }\\;k\\ge\\tau,\\\\ 0, &\\mathrm{otherwise.} \\end{cases}\n\\]\n\\end{Definition}\nIt turns out $\\mathscr{H}_{\\alpha}=\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$, where the equality here means the norm equivalence. \nHence, results established for the Hermite space $\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ can be readily translated to $\\mathscr{H}_{\\alpha}$ up to a constant, which allows us to compare the results on higher order digital nets in \\cite{DILP2018} with ours.\n\nA proof of this equivalence of the norm is outlined in \\cite[1.5.4.~Proposition]{B1998}. \nA more detailed proof of one direction of the equivalence,  $f\\in\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ implying $f\\in\\mathscr{H}_{\\alpha}$ with the same $\\alpha\\in\\mathbb{N}$, is given in \\cite[Lemma~6]{DILP2018}. \nHere, for completeness we prove its converse.\n\\begin{lemma}\nLet $f\\in\\mathscr{H}_{\\alpha}$ with $\\alpha\\in\\mathbb{N}$, then $f\\in\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ with the same smoothness parameter $\\alpha$.\n\\end{lemma}\n\\begin{proof}\nWe first prove the claim for $\\alpha=1$. Assume $f\\in\\mathscr{H}_{1}$. Let \n$F(x):=f(x)\\; (\\rho (x))^{1/2+\\varepsilon}$ and $ G(x):=H_k(x)(\\rho (x))^{1/2-\\varepsilon}$ for $0<\\varepsilon<1/2$.\nWe have\n\\[\\int_{\\bbR}F'(x) \\phi(x) \\mathrm{d} x\n=\n-  \\int_{\\bbR}F(x) \\phi'(x) \\mathrm{d} x,\n\\]\nfor any function $\\phi$ in the space of compactly supported infinitely differentiable functions $C^\\infty _{\\mathrm{c}} (\\bbR)$.\nSince $G$ is in the standard Sobolev space $W^{1,2} (\\bbR)$, \nthere exists a sequence $\\{\\phi_N\\}_{N\\in\\mathbb{N}}\\subset C^\\infty _{\\mathrm{c}} (\\bbR)$ that satisfies\n\\[\n \\|G-\\phi_N\\|^2_{L^2(\\bbR)} +  \\|G'-\\phi'_N\\|^2_{L^2(\\bbR)} = \\|G-\\phi_N\\|^2_{W^{1,2}(\\bbR)} \\le 1/N.\n\\]\nThen, letting $g\\in L^2(\\mathbb{R})$, the Cauchy--Schwarz inequality implies \n$\\int_{\\bbR} |F(x)g(x)| \\mathrm{d} x\n     \\le\n     \\|f\\|_{L^2_\\rho}\\|g\\|_{L^2(\\bbR)}<\\infty$, \nand for\n$F'(x)=f'(x) (\\rho (x))^{1/2+\\varepsilon}-(\\varepsilon+1/2)xf(x)(\\rho (x))^{1/2+\\varepsilon}$ we also have\n\\begin{align*}\n     \\int_{\\bbR} |F'(x)g(x)| \\mathrm{d} x\n     &\\le\n     \\left(\\int_{\\bbR} | f'(x) |^2 \\mathrm{e}^{-x^2/2} \\mathrm{d} x \\right)^{1/2}\\;\n     \\left(\\int_{\\bbR}  |g(x)|  \\mathrm{d} x \\right)^{1/2}\n\\\\\n     &\\kern-1.5cm +\n     \\sup_{t\\in\\mathbb{R}}|(\\varepsilon+1/2)^2t^2 \\mathrm{e}^{-\\varepsilon t^2}|\n     \\left(\\int_{\\bbR} | f(x) |^2 \\mathrm{e}^{-x^2/2} \\mathrm{d} x \\right)^{1/2}\\;\n     \\left(\\int_{\\bbR}  |g(x)|^2 \\mathrm{d} x \\right)^{1/2}\n<\\infty.\n\\end{align*}\nTherefore both $\\langle F,\\cdot\\rangle:=\\int_{\\bbR}F(x) \\cdot \\mathrm{d} x$ and $\\langle F',\\cdot \\rangle:=\\int_{\\bbR}F'(x) \\cdot \\mathrm{d} x$ define a continuous functional on $L^2(\\bbR)$. \nHence, we have $\\int_{\\bbR}F'(x)G(x)\\mathrm{d} x=-\\int_{\\bbR}F(x)G'(x)\\mathrm{d} x$ and thus\n\\begin{align*}\n &\\int_{\\bbR} f'(x) H_k(x) \\rho(x) -(\\varepsilon+1/2)x H_k (x)f(x)\\rho (x) \\mathrm{d} x\n =\\int_{\\bbR} F'(x)G(x) \\mathrm{d} x\\\\\n =&-\\int_{\\bbR} F(x)G'(x) \\mathrm{d} x=-\n \\biggl(\\!\\int_{\\bbR}\\!f(x) \\sqrt{k}H_{k-1}(x) \\rho(x) -(-\\varepsilon+1/2)x H_k (x)f(x)\\rho (x) \\mathrm{d} x\n \\biggr),\n\\end{align*}\nwhich is equivalent to \n\\begin{align}\n  \\int_{\\bbR} f'(x)& H_k(x) \\rho(x)=-\\int_{\\bbR} f(x) \\sqrt{k}H_{k-1}(x) \\rho(x)-xH_k (x)f(x)\\rho (x) \\mathrm{d} x\n  \\label{eq:<f',H>-identity}\n  \\\\\n  &=\n  -\\int_{\\bbR} f(x) (H_k (x)\\rho (x))' \\mathrm{d} x=\n  \\int_{\\bbR} f(x) \\sqrt{k+1}H_{k+1} (x)\\rho (x) \\mathrm{d} x\n  ,\\nonumber\n\\end{align}\nwhere we used $(H_k (x)\\rho (x))'=-\\sqrt{k+1}H_{k+1} (x)\\rho (x)$. Hence we obtain\n\\begin{align*}\n (f',H_k)_{L^2_\\rho} ^2 =(k+1)\\;(f,H_{k+1})_{L^2_\\rho} ^2,\n\\end{align*}\nand\n\\[\n \\sum_{k=0} ^\\infty (k+1)\\;(f,H_{k+1})_{L^2_\\rho} ^2=\\sum_{k=0} ^\\infty (f',H_k)_{L^2_\\rho}^2=\\|f'\\|_{L^2_\\rho} ^2 <\\infty.\n\\]\nThis implies $f\\in\\mathcal{H}^{\\mathrm{Hermite}}_{1}$ since $r_1^{-1}(k)=k+1$.\nFor general $\\tau=1,...,\\alpha$, assuming $f\\in\\mathscr{H}_\\tau$ and by repeating the same argument as above, \nwe have $(f^{(\\tau)},H_{k})_{L^2_\\rho}^2=(f,H_{k+\\tau})_{L^2_\\rho}^2\\prod_{j=1}^{\\tau}(k+j)$ for $k\\geq0$ and thus\n\\begin{align*}\n\\|f^{(\\tau)}\\|_{L^2_\\rho}^2\n&=\n\\sum_{k=0}^{\\infty}\n    (f,H_{k+\\tau})_{L^2_\\rho}^2\\prod_{j=1}^{\\tau}(k+j)\n\\geq\\frac{1}{\\tau^{\\tau}}\\sum_{k=0}^{\\infty}(f,H_{k+\\tau})_{L_{\\rho}^{2}}^{2}(k+\\tau)^{\\tau}\\\\\n&\n=\\frac{1}{\\tau^{\\tau}}\\sum_{k=\\tau}^{\\infty}k^{\\tau}(f,H_{k})_{L_{\\rho}^{2}}^{2}\n.\n\\end{align*}\nHence, observing $\\lim_{k\\to\\infty }r_\\tau(k)\\,k^{\\tau}=1$ (see also \\cite[p.~687]{DILP2018}), we conclude $f\\in \\mathcal{H}^{\\mathrm{Hermite}}_{\\tau}$.\n\\end{proof} \n\n\\section{Matching bounds for Gauss--Hermite quadrature}\\label{sec:GH-bounds}\nIn this section, we prove the sub-optimality of Gauss--Hermite quadrature. \nWe first introduce the following linear quadrature of general form\n\\begin{align}\\label{eq:general_quadrature}\n    Q_n(f)=\\sum_{j=1}^{n}w_{j} f(\\xi_{j})\n\\end{align}\nwith arbitrary $n$ distinct quadrature points on the real line\n\\[ -\\infty<\\xi_1<\\xi_2<\\cdots <\\xi_n<\\infty \\]\nand quadrature weights $w_1,\\ldots,w_n\\in \\mathbb{R}$.\nGauss--Hermite quadrature $Q_{n}^{\\mathrm{GH}}$ is given by the points $(\\xi_{j}^{\\mathrm{GH}})_{j=1,\\dots,n}$ being the roots of $H_n$ and the weights $(w_{j})_{j=1,\\dots,n}$ being  \n$\nw_{j}={1}/{[H_{n}'(\\xi_{j}^{\\mathrm{GH}})]^{2}}\n$,\nsee for example \\cite[Theorem 3.5]{Shen.J_2011_book}.\n\nGiven a quadrature rule $Q_n$, \nit is convenient to introduce the notation\n\\[\n{e}^{\\mathrm{wor}}(Q_{n}, \\mathscr{H}_{\\alpha})\n:=\n\\sup_{0\\neq f\\in \\mathscr{H}_{\\alpha}}\n\\frac{|I(f)-Q_n(f)|}\n    {\\|f\\|_{\\alpha}}\n.\n\\]\nThe quantity ${e}^{\\mathrm{wor}}(Q_{n}, \\mathscr{H}_{\\alpha})$ is commonly referred to as the \\emph{worst-case error} of $Q_n$ in $\\mathscr{H}_{\\alpha}$; see for example \\cite{Dick.J_Kuo_Sloan_2013_ActaNumerica}.\nNow we can state our aim of this section more precisely: we prove the matching lower and upper bounds on ${e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha})$ for Gauss--Hermite quadrature.\n\\subsection{Lower bound}\nWe first derive the following lower bound on ${e}^{\\mathrm{wor}}(Q_{n}, \\mathscr{H}_{\\alpha})$ for the general quadrature \\eqref{eq:general_quadrature}.\n\\begin{lemma}\\label{lem:lower_bound_general}\nLet $\\alpha\\in\\mathbb{N}$.\nFor $n\\geq 2$, let\n\\begin{align}\\label{eq:assum_minimum_distance}\n    \\sigma := \\begin{cases} \\alpha & \\text{if }\\ \\min_{j=1,\\ldots,n-1}(\\xi_{j+1}-\\xi_{j})\\leq 1.\\\\\n    0 & \\text{otherwise.}\\end{cases}\n\\end{align}\nThen, there exists a constant $c_{\\alpha}>0$, which depends only on $\\alpha$, such that the worst-case error ${e}^{\\mathrm{wor}}(Q_n, \\mathscr{H}_{\\alpha})$ of a general function-value based linear quadrature  \\eqref{eq:general_quadrature} in the weighted Sobolev space $\\mathscr{H}_{\\alpha}$ is bounded below by\n\\begin{align*}\n    &{e}^{\\mathrm{wor}}(Q_n, \\mathscr{H}_{\\alpha})\\geq c_{\\alpha}\\min_{i=1,\\dots,n-1}(\\xi_{i+1}-\\xi_{i})^{\\sigma+1/2}\n    \\\\&\n    \\qquad\\quad\\times \\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}\n        \\left(\\sum_{k=1}^{n-1}\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{k}\\xi_{k+1}) \\min(\\xi_{k}^2,\\xi_{k+1}^2)/2}\\right)^{-1/2},\n\\end{align*}\nwhere $\\mathds{1}_{\\geq 0}(x)$ is equal to 1 if $x\\geq 0$ and 0 otherwise.\n\\end{lemma}\n\\begin{proof}\nThe heart of the matter is to construct a function $0\\neq h_{n}\\in \\mathscr{H}_{\\alpha}$ such that $h_{n}(\\xi_{j})=0$ for all $1\\leq j\\leq n$, resulting in $Q_n(h_{n})=0$, and that  $\\|h_n\\|_{\\alpha}$ is small but $I(h_n)$ is large.\nDefine a function $h\\colon \\mathbb{R}\\to \\mathbb{R}$ by\n\n\\[ h(x):=h_n(x)\n:=\\begin{cases} \n    \\displaystyle\n    \\biggl(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\biggr)^{\\!\\!\\alpha}  \\biggl(1-\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\biggr)^{\\!\\!\\alpha} \n        &\\kern-4mm  \n        \\begin{array}{l}\n        \\text{if there exists }j\\!\\in\\! \\{1,\\ldots,n-1\\}\\\\[-2.5pt]\n        \\text{such that }x\\in [\\xi_{j},\\xi_{j+1}],\n        \\end{array}\n        \\\\[10pt]\n    0  &\\kern-4mm\\begin{array}{l}\\text{otherwise.}\\end{array}\n\\end{cases}\n\\]\nThen, $h$ turns out to fulfill our purpose. This type of fooling function used to prove lower bounds for the worst-case error is called \\emph{bump function} (of finite smoothness), in quasi-Monte Carlo theory; see, for instance, \\cite[Section~2.7]{DHP2015}.\n\nFirst we show $h\\in \\mathscr{H}_{\\alpha}$. It follows from\n\\[ \\left(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha}\\left(1-\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha}=\\sum_{\\ell=0}^{\\alpha}(-1)^{\\ell}\\binom{\\alpha}{\\ell}\\left(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha+\\ell}\\]\nthat we have\n\\begin{align*}\n    h^{(\\tau)}(x)=\\frac{1}{(\\xi_{j+1}-\\xi_{j})^{\\tau}}\\sum_{\\ell=0}^{\\alpha}(-1)^{\\ell}\\binom{\\alpha}{\\ell}\\frac{(\\alpha+\\ell)!}{(\\alpha+\\ell-\\tau)!}\\left(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha+\\ell-\\tau}\n\\end{align*}\nfor $\\tau=0,1,\\ldots,\\alpha$ and any $\\xi_{j}<x<\\xi_{j+1}$. \nAs we have $h^{(\\tau)}(\\xi_{j})=h^{(\\tau)}(\\xi_{j+1})=0$ for $\\tau=0,1,\\ldots,\\alpha-1$, the function $h$ is $(\\alpha-1)$-times continuously differentiable. Moreover, $h^{(\\alpha-1)}$ is continuous piecewise polynomial and thus weakly differentiable. Also, noting that $\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1})=0$ only when $\\xi_{j}<0<\\xi_{j+1}$ and otherwise $\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1})=1$, we have\n\\begin{align*}\n    & \\int_{\\xi_{j}}^{\\xi_{j+1}}| h^{(\\tau)}(x)|^2\\rho(x)\\mathrm{d} x \\\\\n    & \\leq \\frac{\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}}{\\sqrt{2\\pi}}\\int_{\\xi_{j}}^{\\xi_{j+1}}| h^{(\\tau)}(x)|^2\\mathrm{d} x \n        \\\\\n    & = \\frac{\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}}{\\sqrt{2\\pi}(\\xi_{j+1}-\\xi_{j})^{2\\tau}} \\\\\n        &{\n            \\times \\kern-3mm\\sum_{\\ell_1,\\ell_2=0}^{\\alpha}\\kern-1.5mm(-1)^{\\ell_1+\\ell_2} \\kern-0.5mm\n            \\binom{\\alpha}{\\ell_1}\\binom{\\alpha}{\\ell_2}\\frac{(\\alpha+\\ell_1)!}{(\\alpha+\\ell_1-\\tau)!}\n                \\frac{(\\alpha+\\ell_2)!}{(\\alpha+\\ell_2-\\tau)!}\n                \\int_{\\xi_{j}}^{\\xi_{j+1}}\\kern-1mm\\Bigl(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\Bigr)^{2(\\alpha-\\tau)+\\ell_1+\\ell_2}\\kern-0.6mm\\mathrm{d} x} \\\\\n    & = \\frac{\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}}{\\sqrt{2\\pi}(\\xi_{j+1}-\\xi_{j})^{2\\tau-1}}\\\\&\n    \\quad{\\times \\sum_{\\ell_1,\\ell_2=0}^{\\alpha}\n        \\frac{(-1)^{\\ell_1+\\ell_2}}{2(\\alpha-\\tau)+\\ell_1+\\ell_2+1}\\binom{\\alpha}{\\ell_1}\\binom{\\alpha}{\\ell_2}\\frac{(\\alpha+\\ell_1)!}{(\\alpha+\\ell_1-\\tau)!}\\frac{(\\alpha+\\ell_2)!}{(\\alpha+\\ell_2-\\tau)!},}\n\\end{align*}\nfor $\\tau=0,1,\\ldots,\\alpha$. The last sum over $\\ell_1$ and $\\ell_2$ does not depend on $j$. \nDenoting this sum by $S_{\\alpha,\\tau}$, we obtain\n\\begin{align*}\n    \\|h\\|_{\\alpha}^2 & = \\sum_{\\tau=0}^{\\alpha}\\int_{\\mathbb{R}}| h^{(\\tau)}(x)|^2 \\rho(x)\\mathrm{d} x = \\sum_{\\tau=0}^{\\alpha}\\sum_{j=1}^{n-1}\\int_{\\xi_{j}}^{\\xi_{j+1}}| h^{(\\tau)}(x)|^2\\rho(x)\\mathrm{d} x \\\\\n    & \\leq \\frac{1}{\\sqrt{2\\pi}}\\sum_{\\tau=0}^{\\alpha}S_{\\alpha,\\tau}\\sum_{j=1}^{n-1}\\frac{\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}}{(\\xi_{j+1}-\\xi_{j})^{2\\tau-1}}\\\\\n    & \\leq\n        \\frac{1}{\\sqrt{2\\pi}}\n        \\sum_{\\tau=0}^{\\alpha}\n            \\frac{S_{\\alpha,\\tau}}\n            {\\min_{1\\le i\\le n-1}(\\xi_{i+1}-\\xi_{i})^{2\\tau-1}}\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}\\\\\n    & \\leq\n    \\frac{1}{\\sqrt{2\\pi}\\min_{1\\le i\\le n-1}(\\xi_{i+1}-\\xi_{i})^{2\\sigma-1}}\\sum_{\\tau=0}^{\\alpha}S_{\\alpha,\\tau}\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{j}\\xi_{j+1}) \\min(\\xi_{j}^2,\\xi_{j+1}^2)/2}<\\infty.\n\\end{align*}\nThis proves $h\\in \\mathscr{H}_{\\alpha}$.\n\nBy definition of $h$, we have $h(\\xi_{j})=0$ for all $j=1,\\ldots,n$, and thus\n\\[ Q_n(h)=0.\\]\nMoreover, we have\n\\begin{align*}\n    I(h) & = \\int_{\\mathbb{R}}h(x)\\rho(x)\\mathrm{d} x = \\sum_{j=1}^{n-1}\\int_{\\xi_{j}}^{\\xi_{j+1}}h(x)\\rho(x)\\mathrm{d} x \\\\\n    & \\geq \\frac{1}{\\sqrt{2\\pi}}\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}\\int_{\\xi_{j}}^{\\xi_{j+1}}\\left(\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha}\\left(1-\\frac{x-\\xi_{j}}{\\xi_{j+1}-\\xi_{j}}\\right)^{\\alpha}\\mathrm{d} x \\\\\n    & = \\frac{1}{\\sqrt{2\\pi}}\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}(\\xi_{j+1}-\\xi_{j})\\int_{0}^{1}x^{\\alpha}\\left(1-x\\right)^{\\alpha}\\mathrm{d} x \\\\\n    & = \\frac{(\\alpha!)^2}{(2\\alpha+1)!\\sqrt{2\\pi}}\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}(\\xi_{j+1}-\\xi_{j}) \\\\\n    & \\geq \\frac{(\\alpha!)^2}{(2\\alpha+1)!\\sqrt{2\\pi}}\n    \\min_{1\\le i\\le n-1}(\\xi_{i+1}-\\xi_{i})\\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}.\n\\end{align*}\n\nUsing the above results, we obtain\n\\begin{align*}\n    {e}^{\\mathrm{wor}}(Q_n,\\mathscr{H}_{\\alpha}) & \n    \\geq \\frac{|I(h)-Q_n(h)|}{\\|h\\|_{\\alpha}}\\\\\n    & \\geq \\frac{(\\alpha!)^2}{(2\\alpha+1)!(2\\pi)^{1/4}}\\left(\\sum_{\\tau=0}^{\\alpha}S_{\\alpha,\\tau}\\right)^{-1/2} \\min_{1\\le i\\le n-1}(\\xi_{i+1}-\\xi_{i})^{\\sigma+1/2}\\\\\n    &{\\quad \\times \\sum_{j=1}^{n-1}\\mathrm{e}^{-\\max(\\xi_{j}^2,\\xi_{j+1}^2)/2}\n    \\left(\\sum_{k=1}^{n-1}\\mathrm{e}^{-\\mathds{1}_{\\geq 0}(\\xi_{k}\\xi_{k+1}) \\min(\\xi_{k}^2,\\xi_{k+1}^2)/2}\\right)^{-1/2}.}\n\\end{align*}\nNow the proof is complete.\n\\end{proof}\nUsing the general lower bound in Lemma~\\ref{lem:lower_bound_general}, we obtain the following lower bound on the worst-case error for Gauss--Hermite quadrature.\n\\begin{theorem}\\label{thm:GH-LB}\nLet $\\alpha\\in\\mathbb{N}$.\nFor any $n\\geq 2$, the worst-case error of the Gauss--Hermite quadrature in the weighted Sobolev space $\\mathscr{H}_{\\alpha}$ is bounded from below as\n\\[ {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha})\\geq C_{\\alpha}n^{-\\alpha/2}\\]\nwith a constant $C_{\\alpha}>0$ that depends on $\\alpha$ but independent of $n$.\n\\end{theorem}\n\\begin{proof}\nLet $\\xi_{j}^{\\mathrm{GH}}$, $j=1,\\dots,n$, be the roots of $H_n$. \nFor any $n\\geq 2$, it holds that\n\\begin{align}\\label{eq:minimum_distance}\n\\frac{\\pi}{\\sqrt{n+1/2}}<\\min_{j=1,\\dots,n-1} (\\xi_{j+1}^{\\mathrm{GH}}-\\xi_{j}^{\\mathrm{GH}})\\leq \\frac{\\sqrt{21/2}}{\\sqrt{n+1/2}};\n\\end{align} \nsee, for instance, \\cite[Eq.~(6.31.22)]{Szego1975}. \nThus, to invoke  Lemma~\\ref{lem:lower_bound_general} we let\n\\[ \\sigma := \\begin{cases} \\alpha & \\text{for }\\ n\\geq 10,\\\\\n    0 & \\text{for }\\ 2\\leq n<10,\\end{cases}\n\\]\nso that the conditions in \\eqref{eq:assum_minimum_distance} are satisfied.\nAlso, each node $\\xi_{j}^{\\mathrm{GH}}$ is bounded below and above as follows; see, for instance, \\cite[Eq.~(6.31.19)]{Szego1975}:\nfor $n$ odd, we have $\\xi_{(n+1)/2}^{\\mathrm{GH}}=0$ with the positive zeros satisfying\n\\begin{align}\\label{eq:node_distribution_odd}\n    \\frac{j\\pi}{\\sqrt{n+1/2}} <\\xi_{(n+1)/2+j}^{\\mathrm{GH}}<\\frac{4j+3}{\\sqrt{n+1/2}}\n    \\quad\\text{for }\\ j=1,\\ldots,(n-1)/2,\n\\end{align}\nand for $n$ even,\n\\begin{align}\\label{eq:node_distribution_even}\n \\frac{(j-1/2)\\pi}{\\sqrt{n+1/2}} <\\xi_{n/2+j}^{\\mathrm{GH}}<\\frac{4j+1}{\\sqrt{n+1/2}}\n    \\quad\\text{for }\\ j=1,\\ldots,n/2,\n\\end{align}\nwith symmetricity $\\xi_{j}^{\\mathrm{GH}} = -\\xi_{n+1-j}^{\\mathrm{GH}}$, $1\\leq j\\leq n$ for $n\\geq 2$ odd and even. \n\nLet $n$ be odd. Using the result in Lemma~\\ref{lem:lower_bound_general}, equations~\\eqref{eq:minimum_distance}, and \\eqref{eq:node_distribution_odd}, together with the symmetricity of the Hermite zeros, we obtain\n\\begin{align*}\n    & {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha}) \\\\\n    & \\geq 2c_{\\alpha}\n        \\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\\,\\sum_{j=1}^{(n-1)/2}\\mathrm{e}^{-(\\xi_{(n+1)/2+j}^{\\mathrm{GH}})^2/2}\n    \\left(2\\sum_{k=1}^{(n-1)/2}\\mathrm{e}^{-(\\xi_{(n+1)/2+k-1}^{\\mathrm{GH}})^2/2}\\right)^{-1/2}\\\\\n& \\geq 2c_{\\alpha}\n        \\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\\,\\sum_{j=1}^{(n-1)/2}\\kern-2mm\\mathrm{e}^{-(4j+3)^2/(2n+1)}\n        \\left(2+2\\kern-2.5mm\\sum_{k=2}^{(n-1)/2}\\kern-2mm\\mathrm{e}^{-\\pi^2(k-1)^2/(2n+1)}\\right)^{-1/2}.\n\\end{align*}\nThe sum over $j$ is further bounded below by\n\\begin{align*}\n    \\sum_{j=1}^{(n-1)/2}\\mathrm{e}^{-(4j+3)^2/(2n+1)} & \\geq \\int_1^{(n-1)/2+1}\\mathrm{e}^{-(4x+3)^2/(2n+1)}\\mathrm{d} x\\\\\n    & = \\frac{\\sqrt{n+1/2}}{4}\\int_{7/\\sqrt{n+1/2}}^{(2n+5)/\\sqrt{n+1/2}}\\mathrm{e}^{-x^2/2}\\mathrm{d} x\\\\\n    & \\geq \\frac{\\sqrt{n+1/2}}{4}\\int_{\\sqrt{14}}^{11\\sqrt{2}/\\sqrt{7}}\\mathrm{e}^{-x^2/2}\\mathrm{d} x\\\\\n    & = \\frac{\\sqrt{\\pi(n+1/2)}}{4\\sqrt{2}}\\left( \\erf(11/\\sqrt{7})-\\erf(\\sqrt{7})\\right),\n\\end{align*}\nwhere $\\erf$ denotes the error function, and the last inequality holds for any odd $n\\geq 3$. \nThe sum over $k$ is further bounded above by\n\\begin{align*}\n    \\sum_{k=2}^{(n-1)/2}\\mathrm{e}^{-\\pi^2(k-1)^2/(2n+1)} & \\leq \\int_1^{(n-1)/2}\\mathrm{e}^{-\\pi^2(x-1)^2/(2n+1)}\\mathrm{d} x \\\\\n    & \\leq \\sqrt{n+1/2} \\int_{0}^{\\infty}\\mathrm{e}^{-\\pi^2x^2/2}\\mathrm{d} x = \\sqrt{\\frac{n+1/2}{2\\pi}}.\n\\end{align*}\nUsing these bounds, we have\n\\begin{align*}\n    {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha}) \n    & \\geq 2c_{\\alpha}\n        \\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\\frac{\\sqrt{\\pi(n+1/2)}}{4\\sqrt{2}}\n        \\left( \\erf(11/\\sqrt{7})-\\erf(\\sqrt{7})\\right)\\\\\n    & \\quad \\times \\left(2+2\\sqrt{\\frac{n+1/2}{2\\pi}}\\right)^{-1/2}\\\\\n    & \\geq c_{\\alpha}\\pi^{\\sigma+1/4} \\frac{\\erf(11/\\sqrt{7})-\\erf(\\sqrt{7})}{2\\sqrt{2}\\sqrt{2+\\sqrt{2}}} \\frac{1}{(n+1/2)^{\\sigma/2}}\\\\\n    & \\geq c_{\\alpha}\\pi^{1/4} \\frac{\\erf(11/\\sqrt{7})-\\erf(\\sqrt{7})}{2^{(\\alpha+5)/2}} \\frac{1}{n^{\\alpha/2}}.\n\\end{align*}\n\nLet $n$ be even. \nAs in the odd case, but now using \\eqref{eq:node_distribution_even} instead of \\eqref{eq:node_distribution_odd}, we obtain\n\\begin{align*}\n    & {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha}) \\\\\n    & \\geq c_{\\alpha}\n        \\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\n        \\\\\n        & \\qquad\\times\n        \\Biggl(\\mathrm{e}^{-(\\xi_{n/2+1}^{\\mathrm{GH}}})^2/2+2\\sum_{j=2}^{n/2}\\mathrm{e}^{-(\\xi_{n/2+j}^{\\mathrm{GH}})^2/2}\\Biggr)\n        \\Biggl(1+2\\sum_{k=2}^{n/2}\\mathrm{e}^{-(\\xi_{n/2+k-1}^{\\mathrm{GH}})^2/2}\\Biggr)^{-1/2}\\\\\n    & \\geq c_{\\alpha}\\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\\\\\n    & \\qquad\\times\n    \\Biggl(\\mathrm{e}^{-5^2/(2n+1)}+2\\sum_{j=2}^{n/2}\\mathrm{e}^{-(4j+1)^2/(2n+1)}\\Biggr)\n    \\Biggl(1+2\\sum_{k=2}^{n/2}\\mathrm{e}^{-\\pi^2(k-3/2)^2/(2n+1)}\\Biggr)^{-1/2}.\n\\end{align*}\nThe sum over $j$ is equal to 0 for $n=2$ and is bounded below by\n\\begin{align*}\n    \\sum_{j=2}^{n/2}&\\mathrm{e}^{-(4j+1)^2/(2n+1)}  \\geq \\int_2^{n/2+1}\\mathrm{e}^{-(4x+1)^2/(2n+1)}\\mathrm{d} x\\\\\n    & = \\frac{\\sqrt{n+1/2}}{4}\\int_{9/\\sqrt{n+1/2}}^{(2n+5)/\\sqrt{n+1/2}}\\mathrm{e}^{-x^2/2}\\mathrm{d} x\\\\\n    & \\geq \\frac{\\sqrt{n+1/2}}{4}\\int_{3\\sqrt{2}}^{13\\sqrt{2}/3}\\mathrm{e}^{-x^2/2}\\mathrm{d} x = \\frac{\\sqrt{\\pi(n+1/2)}}{4\\sqrt{2}}\\left( \\erf(13/3)-\\erf(3)\\right),\n\\end{align*}\nfor $n\\geq 4$. \nNoting that we have $\\mathrm{e}^{-5}\\geq \\sqrt{5\\pi}\\left( \\erf(13/3)-\\erf(3)\\right)/4$, it holds that\n\\[ \\mathrm{e}^{-5^2/(2n+1)}+2\\sum_{j=2}^{n/2}\\mathrm{e}^{-(4j+1)^2/(2n+1)} \\geq \\frac{\\sqrt{\\pi(n+1/2)}}{2\\sqrt{2}}\\left( \\erf(13/3)-\\erf(3)\\right),\\]\nfor any even $n$. \nThe sum over $k$ is again equal to 0 for $n=2$ and is bounded above by\n\\begin{align*}\n    \\sum_{k=2}^{n/2}\\mathrm{e}^{-\\pi^2(k-3/2)^2/(2n+1)} & = \\sum_{k=1}^{n/2-1}\\mathrm{e}^{-\\pi^2(k-1/2)^2/(2n+1)}  \\leq \\int_0^{ n/2-1}\\mathrm{e}^{-\\pi^2(x-1/2)^2/(2n+1)}\\mathrm{d} x \\\\\n    & \\leq \\sqrt{n+1/2} \\int_{-\\infty}^{\\infty}\\mathrm{e}^{-\\pi^2x^2/2}\\mathrm{d} x = \\sqrt{\\frac{2n+1}{\\pi}}.\n\\end{align*}\nIt follows from these bounds on the sums that\n\\begin{align*}\n    {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha}) & \\geq c_{\\alpha}\\left( \\frac{\\pi^2}{n+1/2}\\right)^{\\sigma/2+1/4}\\frac{\\sqrt{\\pi(n+1/2)}}{2\\sqrt{2}}\\\\\n    &\\qquad\\times\n    \\bigl( \\erf(13/3)-\\erf(3)\\bigr)\n        \\left(1+2\\sqrt{\\frac{2n+1}{\\pi}}\\right)^{-1/2} \\\\\n    & \\geq c_{\\alpha}\\pi^{\\sigma+1/4} \\frac{\\erf(13/3)-\\erf(3)}{2\\sqrt{2}\\sqrt{2\\sqrt{2}+2}}\\frac{1}{(n+1/2)^{\\sigma/2}}\\\\\n    & \\geq c_{\\alpha}\\pi^{1/4} \\frac{\\erf(13/3)-\\erf(3)}{2^{(\\alpha+6)/2}}\\frac{1}{n^{\\alpha/2}}.\n\\end{align*}\n\nAltogether, we obtain a lower bound for the worst-case error \n\\[ {e}^{\\mathrm{wor}}(Q_{n}^{\\mathrm{GH}}, \\mathscr{H}_{\\alpha})\\geq C_{\\alpha}n^{-\\alpha/2}\\]\nwith\n\\begin{align*}\n    C_{\\alpha}\n    &=c_{\\alpha}\\pi^{1/4} \\min\\Biggl\\{\\frac{\\erf(11/\\sqrt{7})-\\erf(\\sqrt{7})}{2^{(\\alpha+5)/2}}, \\frac{\\erf(13/3)-\\erf(3)}{2^{(\\alpha+6)/2}}\\Biggr\\}\\\\\n    &=\n    c_{\\alpha}\\pi^{1/4} \\frac{\\erf(13/3)-\\erf(3)}{2^{(\\alpha+6)/2}},\n\\end{align*}\nwhich holds for all $n\\geq 2$.\n\\end{proof}\nThe general lower bound in Lemma~\\ref{lem:lower_bound_general} depends on the set of quadrature points but not on the set of weights. \nBecause the lower bound for Gauss--Hermite quadrature in Theorem~\\ref{thm:GH-LB} is built up on Lemma~\\ref{lem:lower_bound_general}, the sub-optimality of Gauss--Hermite quadrature holds irrespective of the choice of the quadrature weights. \n\nThe proof of Lemma~\\ref{lem:lower_bound_general} in particular indicates that, if the spacing of a node set decreases asymptotically no faster than $1/\\sqrt{n}$, then the corresponding quadrature rule cannot achieve the worst-case error better than $O(n^{-\\alpha/2})$. \nTo elaborate this point, we present the following less tight but more general result, which implies that any function-value based quadrature rule that does not have a quadrature point in $[0,n^{-1/2}]$, say, cannot have a worst-case error better than $O(n^{-\\alpha/2-1/4})$.\n\\begin{corollary}\\label{cor:delta-LB}\nFor $f\\in\\mathscr{H}_{\\alpha}$ with $\\alpha\\in\\mathbb{N}$, \nlet $Q_n(f)$ be a quadrature of the form~\\eqref{eq:general_quadrature}. \nTake a positive number \n$\\delta=\\delta(n)\\in(0,1]$\nsuch that no quadrature\npoint is in $(0,\\delta)$. \n Then, we have\n\\[\n{e}^{\\mathrm{wor}}(Q_n, \\mathscr{H}_{\\alpha})\\geq C_{\\alpha}\\delta^{\\alpha+1/2},\n\\]\nwhere the constant $C_{\\alpha}>0$ is independent of $\\delta$. In\nparticular, if ${Q}_{n}$ does not have any quadrature point in $(0,n^{-r})$, $r>0$, then we have ${e}^{\\mathrm{wor}}(Q_n, \\mathscr{H}_{\\alpha})\\geq C_{\\alpha}n^{-r\\alpha-r/2}$.\n\\end{corollary}\n\\begin{proof}\nConsider a function $f_{\\delta,\\alpha}$ defined by  $f_{\\delta,\\alpha}(x):=(x/\\delta)^{\\alpha}(1-x/\\delta)^{\\alpha}\\mathds{1}_{[0,\\delta]}(x)$,\n$x\\in\\mathbb{R}$. Then, following the proof of Lemma~\\ref{eq:general_quadrature},\nanalogous calculations show that we have  $|I(f_{\\delta,\\alpha})-Q_{n}(f_{\\delta,\\alpha})|=|I(f_{\\delta,\\alpha})|\\geq C_\\alpha\\|f_{\\delta,\\alpha}\\|_{\\alpha}\\delta^{\\alpha+1/2}$ \nfor a constant $C_\\alpha>0$. \nThis completes the proof.\n\\end{proof}\n\nIn passing, we note that the Theroem~\\ref{thm:GH-LB} also gives a lower bound for the interpolation $L^1_{\\rho}$-error. \nGiven a function $f\\colon\\mathbb{R}\\to\\mathbb{R}$, let $\\Lambda_{n}f$\nbe the polynomial interpolant defined by the zeros of $H_{n}$. \nThen,\nlike other interpolatory quadratures, Gauss--Hermite quadrature satisfies\n$\\frac{1}{\\sqrt{2\\pi}}\\int_{\\mathbb{R}}\\Lambda_{n}f(x)\\,\\mathrm{e}^{-x^{2}/2}\\mathrm{d} x=\\sum_{j=1}^{n}w_{j}f(x_{j})$.\nTherefore, \nwith \n$\n \\|f-\\Lambda_{n}f\\|_{L_{\\rho}^{1}(\\mathbb{R})}\n :=\\int_{\\mathbb{R}}\\big|f(x)-\\Lambda_{n}f(x)\\big|\\,\\rho(x)\\mathrm{d}x$\ndenoting the interpolation $L^1_{\\rho}$-error, we have\n\\begin{align*}\n|I(f)-Q^{\\mathrm{GH}}_n(f)|&=\\Bigl|\\int_{\\mathbb{R}}\\big[f(x)-\\Lambda_{n}f(x)\\big]\\,\\rho(x)\\mathrm{d}x\\Bigr|\\leq\n \\|f-\\Lambda_{n}f\\|_{L_{\\rho}^{1}(\\mathbb{R})},\n\\end{align*}\nand thus $\\sup_{0\\not=f\\in \\mathscr{H}_{\\alpha}}\\frac{\\|f-\\Lambda_{n}f\\|_{L_{\\rho}^{1}(\\mathbb{R})}}{\\|f\\|_\\alpha}\\geq C_\\alpha n^{-\\alpha/2}$.\n\n\\subsection{Upper bound}\nIn the previous section, we showed a lower bound for the worst-case error of the rate $n^{-\\alpha/2}$. \nA matching upper bound has been shown by Mastroianni and Monegato, the result of whom we adapt to our setting. \n\\begin{proposition}[\\cite{MM1994}]\nLet $\\alpha\\in\\mathbb{N}$ and $n\\in\\mathbb{N}$. For  $f\\in\\mathscr{H}_{\\alpha}$, let $Q_{n}^{\\mathrm{GH}}(f)$ be the Gauss--Hermite approximation to $I(f)$. Then, we have\n\\[\n|I(f)-Q_{n}^{\\mathrm{GH}}(f)|\\leq C n^{-\\alpha/2}\\|f\\|_{\\alpha},\n\\]\nwhere $C>0$ is a constant independent of $n$ and $f$.\n\\end{proposition}\n\\begin{proof}\nFirst, $f\\in\\mathscr{H}_{\\alpha}$ implies $f^{(\\tau)}\\in W^{1,2}_{\\mathrm{loc}}(\\mathbb{R})$ and thus $f^{(\\tau)}$ admits a locally absolutely continuous representative for $\\tau=0,\\dots,\\alpha-1$. Moreover, \nfrom $f\\in\\mathscr{H}_{\\alpha}$, for any $0<\\varepsilon<1/2$ we have\n\\[\n\\int_{\\mathbb{R}}\n\t\\mathrm{e}^{\\varepsilon x^2/2}|f^{(\\alpha)}(x)|\\mathrm{e}^{-x^2/2}\\mathrm{d}x\n    \\leq \n    \\Bigl(\n    \t\\int_{\\mathbb{R}}\n\t\t|f^{(\\alpha)}(x)|^2\\mathrm{e}^{-x^2/2}\n        \\mathrm{d}x\\Bigr)^{1/2}\n    \\Bigl(\n    \t\\int_{\\mathbb{R}}\n    \t\\mathrm{e}^{-x^2(1/2-\\varepsilon)}\n     \\mathrm{d}x\\Bigr)^{1/2}<\\infty.\n\\]\nThus, from \\cite[Theorem 2]{MM1994}, the statement follows.\n\\end{proof} \\section{Optimality of the trapezoidal rule} \\label{sec:trapez}\nIn this section, we prove the optimality of trapezoidal rules in $\\mathscr{H}_\\alpha$.\nMore precisely, we consider the following quadrature with  $n$ equispaced points\n\\begin{align}\nQ_{n,T}^{{*}}(g):=\\frac{2T}{n}\\sum_{j=0}^{n-1}g(\\xi_j^*), \\quad\\text{with}\\ \\xi_j^*:=\\frac{2T}{n}j-T,\\ j=0,\\dots,n-1.\\label{eq:def-trap}\n\\end{align}\nHere, $T>0$ is a parameter that controls the cut-off of the integration domain from $\\mathbb{R}$ to $[-T,T]$. \n\nWe call $Q_{n,T}^*$ a \\textit{trapezoidal rule}:\nindeed, for $n$ even, $Q_{n,T}^{*}(g)$ is nothing but the standard truncated trapezoidal rule for functions on $\\mathbb{R}$ with mesh size $\\Delta x$\n\\[\nQ^*_{n,T}(g)=\\Delta x \\sum_{\\ell=-n/2}^{n/2-1}g(\\ell\\, \\Delta x),\n\\]\nwith $\\Delta x=2T/n$, while for $n$ odd we have the trapezoidal rule only shifted by $-\\frac12\\Delta x$:\n\\[\nQ^*_{n,T}(g)=\\Delta x \\sum_{\\ell=-(n-1)/2}^{(n-1)/2}g\\Bigl(\\bigl(\\ell-\\frac12\\bigr)\\Delta x\\Bigr).\n\\]\n\nOur proof strategy is based on the approach by Nuyens and Suzuki \\cite{NS2021}, where a multidimensional integration problem with respect to the Lebesgue measure using a quasi-Monte Carlo method called rank-$1$ lattice rule was considered. \nFollowing \\cite{NS2021}, we consider the bound\n\\begin{align}\n\\left|\n    \\int_{\\bbR} g(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(g)\n  \\right|\n  \\le\n  \\biggl|\\int_{\\bbR} g(x) \\mathrm{d} x -\\int_{-T}^T g(x) \\mathrm{d} x  \\biggr| \n  +\n  \\biggl|\\int_{-T}^T g(x) \\mathrm{d} x - Q_{n,T}^* (g) \\biggr|.\n  \\label{eq:error-decomp-g}\n\\end{align}\nWe will let $g=f\\rho$ with $f\\in\\mathscr{H}_\\alpha$ later in Theorem~\\ref{thm:trape-opt}.\n\nThe first term of the right hand side in \\eqref{eq:error-decomp-g} can be bounded as in \\cite[Proposition~8]{NS2021}; \nwe provide a proof adapted to our setting below in Proposition~\\ref{prop:trape}. \nTo bound the second term, we now derive what corresponds to \\cite[Lemma~5 and Proposition~7]{NS2021}. \n\n\\begin{lemma}\n\\label{lem:trap-error-inside}\nLet $T>0$, $\\alpha\\in\\mathbb{N}$, and $n\\in\\mathbb{N}$ be given.\nSuppose that $g^{(\\tau)}\\colon\\mathbb{R}\\to\\mathbb{R}$ is absolutely continuous on any compact interval for $\\tau=0,\\dots,\\alpha-1$, \nand that $g^{(\\alpha)}$ is in $L^2(\\bbR)$.\nSuppose further that $g$ satisfies\n\\begin{equation}\n\\|g\\|_{\\alpha,[-T,T]}\n:=\n\\left( \\left(\\sum_{\\tau=0}^{\\alpha-1}  \\left(\\int_{-T}^T g^{(\\tau)}(x) \\mathrm{d} x\\right)^2 +\\int_{-T}^T | g^{(\\alpha)}(x) |^2\\mathrm{d} x \\right)\n\\right)^{1/2}\n< \\infty\n\\end{equation}\nand\n\\begin{align}\n \\|g\\|_{\\alpha,\\mathrm{decay}} \n :=\n \\sup_{\\substack{x\\in\\bbR \\\\ \\tau\\in\\{0,\\ldots,\\alpha-1\\}}}\n \\left| \\mathrm{e}^{(1-\\varepsilon)x^2/2} \\, g^{(\\tau)}(x) \\right|\n <\n \\infty\n ,\\quad \\text{for some  $\\varepsilon\\in(0,1)$}.\n \\label{eq:decay-cond}\n\\end{align}\nThen the error of the $n$-point trapezoidal rule on the interval $[-T,T]$ defined in \\eqref{eq:def-trap} is bounded by \n\\begin{align}\n\\left|\n    \\int_{-T} ^T g(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(g)\n  \\right| \n   &\\le\n   C_\\alpha\\|g\\|_{\\alpha,[-T,T]} T^{\\alpha+1/2} \\frac{1}{n^{\\alpha}} \\nonumber\n   \\\\&\\quad+\n   \\alpha \\max\\{1,(2T)^{\\alpha-1}\\}\\|g\\|_{\\alpha,\\mathrm{decay}}\\mathrm{e}^{-(1-\\varepsilon)T^2/2}\n  ,\n\\end{align}\nwith $C_\\alpha :=2\\sqrt{\\zeta(2\\alpha)}/\\pi^\\alpha$, where $\\zeta(2\\alpha):=\\sum_{m=1}^{\\infty}m^{-2\\alpha}<\\infty$.\n\\end{lemma}\n\\begin{proof}\nWith a suitable auxiliary function $G=G^{[-T,T]}$ periodic on $[-T,T]$ satisfying $\\int_{-T}^T  G(x) \\mathrm{d} x=\\int_{-T}^T g(x) \\mathrm{d} x$, we consider the following bound:\n\\begin{align}\n     \\left|\n    \\int_{-T} ^T g(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(g)\n  \\right|\n  \\le\n       \\left|\n\\int_{-T} ^T G(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(G)\n  \\right|\n  +\n         \\left|\nQ_{n,T}^*(g-G) \n  \\right|\n  .\\label{eq:[-T,T]-err-decomp}\n\\end{align}\nChoosing\n\\[\nG(x):=g(x)-\\sum_{\\tau=1}^{\\alpha}\\frac{B^{[-T,T]}_{\\tau}(x)}{\\tau!}\\left(\\int_{-T}^T g^{(\\tau)}(s) \\mathrm{d} s\\right)\\text{ for }x\\in [-T-\\delta,T+\\delta],\n\\]\nfor an arbitrarily fixed small $\\delta\\in (0,1)$ turns out to be convenient, as we now explain. Here, \n$B^{[-T,T]}_{\\tau}(x)$ is the scaled Bernoulli polynomial of degree $\\tau$ on $[-T,T]$, namely\n\\[\nB^{[-T,T]}_{\\tau}(x) = (2T)^{\\tau-1} B_{\\tau}\\left(\\frac{x+T}{2T}\\right)\n\\]with $B_{\\tau}$ being the standard Bernoulli polynomial of degree $\\tau$. \nWe have $\\int_{-T}^T  G(x) \\mathrm{d} x=\\int_{-T}^T g(x) \\mathrm{d} x$ by simply noticing that $\\int_0^1 B_\\tau(x)\\mathrm{d} x =0$ for $\\tau\\ge 1$.\n\nThe function $G$ is ($\\alpha-1$)-times differentiable on \n$(-T-\\delta,T+\\delta)$ \nwith $G^{(\\alpha-1)}$ being absolutely continuous on $[-T,T]$. Moreover, we have\n\\begin{align*}\n&\\int_{-T}^{T} G^{(\\tau)}(x)\\mathrm{d} x =\\int_{-T}^{T} g^{(\\tau)}(x)\\mathrm{d} x - \\left(\\int_{-T}^{T} B^{[-T,T]}_{0}(x) \\mathrm{d} x \\right)\\left(\\int_{-T}^{T} g^{(\\tau)}(s)\\mathrm{d} s\\right)  =0 ,\n\\end{align*}\nfor $\\tau=1,\\ldots,\\alpha$, and thus the fundamental theorem of calculus tells us\n\\begin{align*}\n&G^ {(\\tau)}(-T)=G^ {(\\tau)}(T),\\quad \\text{for }\\  \\tau=0,\\ldots,\\alpha-1.\n\\end{align*}\n\nThese properties of $G$ imply the following two Fourier series representations. \nFirst, from the periodicity and the absolute continuity of $G$ on $[-T,T]$, \nwe have the pointwise-convergent Fourier series expansion\n\\[\nG(x)= \\sum_{m\\in\\mathbb{Z}} \\widehat{G}(m) \\phi^{[-T,T]}_{m}(x), \n\\]\nwhere $\\phi^{[-T,T]}_{m}(x):=\\exp(\\frac{2\\pi \\mathrm{i} m(x+T)}{2T})/\\sqrt{2T}$,  $m\\in\\mathbb{Z}$ are the orthonormal Fourier basis on $L^2([-T,T])$ and \n$\\widehat{G}(m):=\\int_{-T}^T G(x) \\exp(\\frac{-2\\pi \\mathrm{i} m(x+T)}{2T})/\\sqrt{2T}\\, \\mathrm{d} x$, $m\\in\\mathbb{Z}$\n are the Fourier coefficients.\nSecond, \nfrom the square integrability of $G^{(\\alpha)}$, we have the $L^2$-convergent Fourier series representation\n\n\\begin{align*}\n    \\left(G(x)\\right)^{(\\alpha)}\n    &=\\sum_{m\\in\\mathbb{Z}} \\widehat{G^{(\\alpha)}}(m) \\phi^{[-T,T]}_{m}(x)=\\sum_{m\\in\\mathbb{Z}} \\left(\\frac{2\\pi \\mathrm{i} m}{2T}\\right)^{\\alpha} \\widehat{G}(m) \\phi^{[-T,T]}_{m}(x),\n\\end{align*}\nwhere in the second equality we repeatedly used the integration by parts.\n\nUsing these representations, we obtain\n\\begin{align}\n   \\biggl|\n   Q_{n,T}^*(G)&\n   -\n    \\int_{-T}^T G(x) \\mathrm{d} x\n    \\biggr|\\notag\n=\n   \\biggl|\n    \\frac{2T}{n} \\sum_{j=0}^{n-1}\\sum_{m\\in\\mathbb{Z}} \\widehat{G}(m) \\phi^{[-T,T]}_{m}(\\xi^*_j)\n    -\n    \\sqrt{2T}\\widehat{G}(0)\n\\biggr|\n\\\\\n&=\n \\Biggl|\n \\sqrt{2T}\\sum_{m\\in\\mathbb{Z}\\setminus\\{0\\}} \\widehat{G}(mn) \n \\Biggr|\\notag\n\\\\\n &\\le\n \\sqrt{2T} \\Biggl(\n \\sum_{m\\in\\mathbb{Z}\\setminus\\{0\\}} |\\widehat{G}(mn)|^2 \\left(\\frac{2\\pi m n}{2T}\\right)^{2\\alpha} \n \\Biggr)^{1/2}\n \\Biggl(\n \\sum_{m\\in\\mathbb{Z}\\setminus\\{0\\}}  \\left(\\frac{2T}{2\\pi m n}\\right)^{2\\alpha} \n \\Biggr)^{1/2}\\notag\n \\\\\n &\\le\n \\sqrt{2T}\\|G\\|_{\\alpha,[-T,T]}\n \\sqrt{2\\zeta(2\\alpha)}\n \\left(\\frac{T}{\\pi n}\\right)^{\\alpha} \n =:C_\\alpha \\|G\\|_{\\alpha,[-T,T]} T^{\\alpha+1/2} \\frac{1}{n^\\alpha},\\label{eq:bd-in-G}\n\\end{align}\nwhere in the first to second lines we used the pointwise convergence of the series and\\begin{align*}\n    \\frac{2T}{n} \\sum_{j=0}^{n-1} \\phi^{[-T,T]}_{m}(\\xi^*_j)\n=\n\\frac{\\sqrt{2T}}{n} \\sum_{j=0}^{n-1} \\exp(2\\pi\\mathrm{i}\\:\\! m j/n)\n=\n\\begin{cases}\n\\sqrt{2T} & \\text{if  } m\\equiv 0 \\pmod{n},\\\\\n0 & \\text{otherwise},\n\\end{cases}\n\\end{align*}\nwhile in the fourth line we used the Parseval identity.\nThe equation \\eqref{eq:bd-in-G} is further bounded by $C_\\alpha \\|g\\|_{\\alpha,[-T,T]} T^{\\alpha+1/2}\\frac{1}{n^\\alpha}$ since\n\\begin{align*}\n\\|G\\|_{\\alpha,[-T,T]}&=\\left(  \\left(\\int_{-T}^T G(x) \\mathrm{d} x\\right)^2 +\\int_{-T}^T | \\left(G\\right)^{(\\alpha)}(x) |^2\\mathrm{d} x\n\\right)^{1/2}\n\\\\\n&=\n\\left(  \\left(\\int_{-T}^T  g(x) \\mathrm{d} x\\right)^2 +\\int_{-T}^T  | g^{(\\alpha)}(x)|^2 \\mathrm{d} x -\\frac{1}{2T}\\left(\\int_{-T}^T  g^{(\\alpha)}(y)\\mathrm{d} y \\right)^2\n\\right)^{1/2}\n\\\\\n&\\le\n\\|g\\|_{\\alpha,[-T,T]}.\n\\end{align*}\nNow we bound the the second term of the right hand side in \\eqref{eq:[-T,T]-err-decomp}:\n\\begin{align*}\n    \\left|\n Q_{n,T}^*(g-G) \n  \\right|\n  &=\n  \\Biggl|\\frac{1}{n}\\sum_{j=0}^{n-1}\\sum_{\\tau=1}^{\\alpha}\\frac{B^{[-T,T]}_{\\tau}(\\xi^*_j)}{\\tau!}\n    \\biggl(\\int_{-T}^T g^{(\\tau)}(s) \\mathrm{d} s\n    \\biggr)\\Biggr|\n  \\\\\n  &\\le\n  \\sum_{\\tau=1}^{\\alpha}\n    \\Bigg|\n        \\frac{1}{n}\\sum_{j=0}^{n-1}\\frac{B^{[-T,T]}_{\\tau}(\\xi^*_j)}{\\tau!}\n    \\Bigg|\n    \\bigl| g^{(\\tau-1)}(T)-g^{(\\tau-1)}(-T)\\bigr|\n  \\\\\n  &\\le\n   \\sum_{\\tau=1}^{\\alpha}\\frac{(2T)^{\\tau-1}}{2} (2 \\|g\\|_{\\alpha,\\mathrm{decay}})\\,\\mathrm{e}^{-(1-\\varepsilon)T^2/2}\n   \\\\\n   &\\le\n   \\alpha \\max\\{1,(2T)^{\\alpha-1}\\}\\|g\\|_{\\alpha,\\mathrm{decay}}\\,\\mathrm{e}^{-(1-\\varepsilon)T^2/2},\n\\end{align*} \nwhere in the penultimate line we used $|\\frac{B^{[-T,T]}_{\\tau}(x)}{\\tau!}|\\le \\frac{(2T)^{\\tau-1}}{2}$ for $x\\in[-T,T]$; see~\\cite[Equation~(6)]{NS2021} or~ \\cite{L1940}. \nTogether with \\eqref{eq:bd-in-G}, the statement follows.\n\\end{proof}\n\nNow, what remains in the bound~\\eqref{eq:error-decomp-g} is the error due to chopping the real line to the interval $[-T,T]$. \nThe following result tells us how to choose $T$ to obtain a total error bounded by $\\mathcal{O}(n^{-\\alpha})$ up to a logarithmic factor. \n\\begin{proposition}\n\\label{prop:trape}\nLet $\\alpha\\in\\mathbb{N}$. \nSuppose that the function $g^{(\\tau)}\\colon\\mathbb{R}\\to\\mathbb{R}$ is absolutely continuous on any compact interval for $\\tau=0,\\dots,\\alpha-1$, \nand that $g^{(\\alpha)}$ is in $L^2(\\bbR)$.\nSuppose further that $g$ satisfies\n\\begin{align}\n\\|g\\|_{\\alpha}^*\n:=\n\\sup_{\\substack{I\\subset\\bbR\\\\|I|<\\infty}} \\|g\\|_{\\alpha,I}\n:=\n\\sup_{\\substack{I\\subset\\bbR\\\\|I|<\\infty}}\\left( \\left(\\sum_{\\tau=0}^{\\alpha-1}  \\left(\\int_I g^{(\\tau)}(x) \\mathrm{d} x\\right)^2 +\\int_I | g^{(\\alpha)}(x) |^2\\mathrm{d} x \\right)\n\\right)^{1/2}\n< \\infty\n\\label{eq:def-alpha-star}\n\\end{align}\nand\n\\begin{align}\n \\|g\\|_{\\alpha,\\mathrm{decay}} \n :=\n \\sup_{\\substack{x\\in\\bbR \\\\ \\tau\\in\\{0,\\ldots,\\alpha-1\\}}}\n \\left| \\mathrm{e}^{(1-\\varepsilon)x^2/2} \\, g^{(\\tau)}(x) \\right|\n <\n \\infty\n ,\\qquad\\text{for some $\\varepsilon\\in(0,1)$}.\\label{eq:decay-cond-proptrap}\n\\end{align}\nThen, for any integer $n\\ge2$, the error for the $n$-point trapezoidal rule $Q^*_{n,T}$ as in \\eqref{eq:def-trap} with the cut-off interval $[-T,T]$ given by\n\\begin{align}\n  T\n  &=\n  \\sqrt{\\frac{2}{(1-\\varepsilon)}\\alpha \\ln(n)}\n  ,\\label{eq:def-T}\n\\end{align}\ncan be bounded by\n\\begin{align}\\label{eq:tr-bound}\n  \\left|\n    \\int_{\\bbR} g(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(g)\n  \\right|\n  \\le\n  C \\, \\left(\\|g\\|_{\\alpha}^* + \\|g\\|_{\\alpha,\\mathrm{decay}}\\right) \\,\n  \\frac{(\\ln n)^{ (\\alpha/2 + 1/4) \\, }}{n^\\alpha}\n  ,\n\\end{align}\nwhere the constant $C$ is independent of $n$ and $g$ but depends on $\\alpha$ and $\\varepsilon$.\n\\end{proposition}\n\\begin{proof}\n\nConsider the bound \\eqref{eq:error-decomp-g}.\nThe error due to cutting off the integration domain is bounded by \n\\begin{align*}\n\\left|\\int_{\\bbR} g(x) \\mathrm{d} x -\\int_{-T}^T g(x) \\mathrm{d} x  \\right| \n&\\le\n 2\\|g\\|_{\\alpha,\\mathrm{decay}} \\int_T^\\infty \\mathrm{e}^{-(1-\\varepsilon)x^2/2} \\mathrm{d} x\n \\\\\n &\\le\n \\frac{2\\|g\\|_{\\alpha,\\mathrm{decay}}}{(1-\\varepsilon)T}\\int_T^\\infty (1-\\varepsilon) x \\mathrm{e}^{-(1-\\varepsilon)x^2/2} \\mathrm{d} x\n \\\\&=\n \\frac{2\\|g\\|_{\\alpha,\\mathrm{decay}}}{(1-\\varepsilon)T} \\mathrm{e}^{-(1-\\varepsilon)T^2/2}\n=\n \\frac{\\sqrt{2}\\|g\\|_{\\alpha,\\mathrm{decay}}}{\\sqrt{\\alpha(1-\\varepsilon)}}n^{-\\alpha} (\\ln(n))^{-1/2}.\n\\end{align*}\nNoting that $n\\ge2$ and \\eqref{eq:def-T} imply $2T>1$ so that $\\max\\{1,(2T)^{\\alpha-1}\\}=(2T)^{\\alpha-1}$,  from Lemma~\\ref{lem:trap-error-inside} we have\n\\begin{align*}\n\\left|\n   \\int_{\\bbR} g(x) \\mathrm{d} x\n    -\n   Q_{n,T}^*(g)\n\\right|\n  &\\le\n    \\frac{\\sqrt{2}\\|g\\|_{\\alpha,\\mathrm{decay}}}{\\sqrt{\\alpha(1-\\varepsilon)}} (\\ln(n))^{-1/2}\\, n^{-\\alpha}\n  \\\\\n  &\\qquad\\quad+\n  C_1 \\, \\|g\\|_{\\alpha}^*\n    \\,\n    (\\ln(n))^{(\\alpha/2+1/4)} \\,\n    n^{-\\alpha}\n  \\\\\n  &\\qquad\\quad+\n  C_2\n    \\,\n    \\|g\\|_{\\alpha,\\mathrm{decay}}\n     \\, (\\ln(n))^{(\\alpha/2-1/2) }\n    \\,\n    n^{-\\alpha}\\\\\n    &\\le\n     C \\, \\left(\\|g\\|_{\\alpha}^* + \\|g\\|_{\\alpha,\\mathrm{decay}}\\right) \\,\n  (\\ln n)^{ (\\alpha/2 + 1/4)} \\, n^{-\\alpha},\n\\end{align*}\nwhere the constants $C_1$, $C_2$ and $C$ are independent of $n$ and $g$ but depend on $\\alpha$ and $\\varepsilon$. Thus the claim is proved.\n\\iffalse\n\\begin{align*}\n\\left|\n    \\int_{\\bbR} g(x) \\mathrm{d} x\n    -\nQ_{n,T}^*(g)\n  \\right|\n  &\\le\n    C_1 \\, \\|g\\|_{\\alpha,\\mathrm{decay}} \\, \\max\\{1, ((1-\\varepsilon) \\, T^2/2)^{1/2}\\} \\, \\frac{2\\exp(-(1-\\varepsilon) \\, T^2/2)}{(1-\\varepsilon) \\, T^2}\n  \\\\\n  &\\qquad\\quad+\n  C_2 \\, \\|g\\|_{\\alpha}^*\n    \\,\n    \\max\\{1, 2T\\}^{(\\alpha+1/2)} \\,\n    \\frac{1}{n^\\alpha}\n  \\\\\n  &\\qquad\\quad+\n  C_3\n    \\,\n    \\|g\\|_{\\alpha,\\mathrm{decay}}\n     \\, \\max\\{1, 2T\\}^{(\\alpha-1) }\n    \\,\n    \\exp(-(1-\\varepsilon) T^2/2) \\\\\n    &\\le\n     C \\, \\left(\\|g\\|_{\\alpha}^* + \\|g\\|_{\\alpha,\\mathrm{decay}}\\right) \\,\n  \\frac{(\\ln n)^{ (\\alpha/2 + 1/4) \\, }}{n^\\alpha}.\n\\end{align*}\n\\fi\n\\end{proof}\n\\begin{remark}\nThe result \\cite[Theorem~2]{NS2021} by Nuyens and Suzuki obtained for a class of quasi-Monte Carlo methods called good lattice rules can be seen as a multidimensional counterpart of Proposition~\\ref{prop:trape}.\nIndeed, it can be checked that the trapezoidal rule is a good lattice, and thus under the same assumption as Proposition~\\ref{prop:trape}, the result therein is immediately applicable to the trapezoidal rule. However, we obtained a better bound by exploiting our one-dimensional setting in Proposition~\\ref{prop:trape}. \nCompare this result with \\cite[Theorem~2]{NS2021} with the parameters therein being $d=1$, $\\beta=(1-\\varepsilon)/2$ and $p=q=2$ to see the improvement.\n\\end{remark}\n\n\nOur results offer several insights to interpret results available in  the literature.\n\n\\begin{sloppypar}\nIn the context of spectral methods, Boyd \\cite{Boyd.JP_2001_book,Boyd.JP_2009} pointed out that the Gauss--Hermite points are distributed roughly uniformly over the interval $[O(-n^{1/2}),O(n^{1/2})]$, and thus, the total number of point being $n$, the spacing between adjacent points decreases only as $O(n^{-1/2})$; see for example \\cite[Chapter~17]{Boyd.JP_2001_book} and \\cite[Fig.~6]{Boyd.JP_2009}.\nThe proof of Theorem~\\ref{thm:GH-LB} (see also Corollary~\\ref{cor:delta-LB}) shows that it is this  slow decrease of the spacing that causes the sub-optimal convergence rate.\n\\end{sloppypar}\n\nIn \\cite[Section 5]{T2021}, Trefethen compared Gausss--Hermite quadrature and various quadrature formulas, including the trapezoidal rule. \nAlthough the focus there was analytic integrands, \nthe author also discusses the nonanalytic case. \nOn page 142, he seems to have reasoned that \n\\footnote{``The ratio increases to nearly order $n^{1/2}$ for nonanalytic functions $f$, where intervals growing just logarithmically rather than algebraically with $n$ are appropriate for balancing domain-truncation and discretization errors.'' \\cite[p.\\ 142]{T2021}.}\nfor the nonanalytic functions on $\\mathbb{R}$ decaying at a suitable rate (presumably at the rate  $\\exp(-x^2)$ as $x\\to\\infty$ including derivatives)  \nthe right choice of the cut-off of the domain that \nbalances domain-truncation and quadrature errors\nshould be logarithmic in $n$,\nwhile the Gauss--Hermite rule distributes quadrature points to unnecessarily wide intervals  $[\\mathcal{O}(-n^{1/2})), \\mathcal{O}(n^{1/2}))]$. \nProposition~\\ref{prop:trape} supports this point for the trapezoidal rule. \nIndeed, \nunder the exponential decay condition \\eqref{eq:decay-cond-proptrap}, \nwe cut off the integration domain logarithmically \\eqref{eq:tr-bound}, and we achieve the optimal rate  $\\mathcal{O}(n^{-\\alpha})$ up to a logarithmic factor. \nNote, however, that for polynomially decaying finitely smooth functions, we expect the right choice of the domain cut-off to grow algebraically; see \\cite[Theorem 2 (ii)]{NS2021} for a related result.\n\nIn Proposition~\\ref{prop:trape}, the choice of the cut-off interval \\eqref{eq:def-T} requires the smoothness parameter $\\alpha$, which might not be known in practice. \nReplacing $\\alpha$ in \\eqref{eq:def-T} with any slowly increasing function $\\gamma(n)$, such as $\\max\\{\\ln(\\ln(n)),0\\}$, yields a less tight bound for the trapezoidal rule, but with an $\\alpha$-free construction, still achieving the optimal rate up to a factor of $(\\gamma(n)\\ln n)^{ (\\alpha/2 + 1/4)}$.\n\\begin{corollary}\\label{cor:alpha-free}\nSuppose that assumptions in Proposition~\\ref{prop:trape} are satisfied. Let $\\gamma(n)\\colon\\mathbb{N}\\to[0,\\infty)$\nbe a non-decreasing function satisfying  $\\lim_{n\\to\\infty}\\gamma(n)=\\infty$.\nThen, for any integer $n\\geq\\gamma^{-1}(\\alpha):=\\min\\{m\\in\\mathbb{N}\\mid\\gamma(m)\\geq\\alpha\\}$,\nthe error for $Q_{n,\\tilde{T}}^{*}$ as in~\\eqref{eq:def-trap} with \n\\[\n\\tilde{T}=\\sqrt{\\frac{2}{(1-\\varepsilon)}\\gamma(n)\\ln(n)}\n\\]\ncan be bounded by \n\\begin{align}\\left|\n    \\int_{\\bbR} g(x) \\mathrm{d} x\n    -\nQ_{n,\\tilde{T}}^*(g)\n  \\right|\n  \\le\n  C \\, \\left(\\|g\\|_{\\alpha}^* + \\|g\\|_{\\alpha,\\mathrm{decay}}\\right) \\,\n  \\frac{(\\gamma(n)\\ln n)^{ (\\alpha/2 + 1/4) \\, }}{n^\\alpha}\n,\n\\end{align}\nwhere the constant $C$ is independent of $n$ and $g$ but depends on $\\alpha$ and $\\varepsilon$.\n\\end{corollary}\n\nNow we are going to show that Proposition~\\ref{prop:trape} is applicable to the weighted Sobolev space $\\mathscr{H}_{\\alpha}$.\nIn view of the optimal rate in \\cite[Theorem 1]{DILP2018} for the Hermite space $\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ and the characterisation of $\\mathcal{H}^{\\mathrm{Hermite}}_{\\alpha}$ with $\\mathscr{H}_{\\alpha}$ discussed in Section~\\ref{sec:space}, the resulting rate below establishes the optimality, up to a logarithmic factor, of our trapezoidal rule.\n\\begin{theorem}\\label{thm:trape-opt}\nFix $\\varepsilon\\in(1/2,1)$ arbitrarily.\nFor $f\\in\\mathscr{H}_{\\alpha}$ with $\\alpha\\in\\mathbb{N}$, consider $Q^*_{n,T}(f\\rho)$ as in \\eqref{eq:def-trap} with\n  $T\n  =\n  \\sqrt{\\frac{2}{1-\\varepsilon}\\alpha \\ln(n)}$. Then, we have\n\\begin{align*}\n  |\n    I(f)\n    -\n    Q_{n,T}^*(f\\rho)\n  |\n  \\le\n  C  \\|f\\|_{\\alpha}  \n  \\frac{(\\ln n)^{ (\\alpha/2 + 1/4) \\, }}{n^\\alpha}\n\\end{align*}\nfor any integer $n\\ge 2$, \nwhere the constant $C$ is independent of $n$ and $f$ but depends on $\\alpha$ and $\\varepsilon$.\n\\end{theorem}\n\\begin{proof}\nLet $g:=f\\rho$. In view of Proposition~\\ref{prop:trape}, it suffices to show $\\|g\\|_{\\alpha}^* + \\|g\\|_{\\alpha,\\mathrm{decay}}\\leq C\\|f\\|_\\alpha$ for some constant $C>0$, where $\\|g\\|_{\\alpha}^*$ and $\\|g\\|_{\\alpha,\\mathrm{decay}}$ are as in \n \\eqref{eq:def-alpha-star} and \\eqref{eq:decay-cond-proptrap} with $\\varepsilon\\in(1/2,1)$, respectively. \n\nWe first show $\\|g\\|_{\\alpha}^*\\leq C \\|f\\|_{\\alpha}$.\nWe have\n\\[\n(\\|g\\|^*_\\alpha)^2 \n\\le\n\\sum_{\\tau=0}^{\\alpha-1}  \\Bigl(\\int_\\bbR |g^{(\\tau)}(x)| \\mathrm{d} x\\Bigr)^2 +\\int_\\bbR | g^{(\\alpha)}(x) |^2\\mathrm{d} x,\n\\]\nbut using \\eqref{eq:Hermite-def} and the chain rule, \nfor $\\tau=0,\\ldots,\\alpha-1$ we have\n\\begin{align*}\n\\|g^{(\\tau)}&\\|_{L^1(\\bbR)}\n\\le\n\\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\| f^{(\\tau-\\ell)}(x)\\rho^{(\\ell)}(x) \\|_{L^1(\\bbR)}\n\\\\\n&=\n\\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\left(\\int_{\\bbR} \\left|\\rho(x) f^{(\\tau-\\ell)}(x)\\sqrt{\\ell!}(-1)^{\\ell} H_{\\ell}(x) \\right| \\mathrm{d} x\\right)\n    \\\\\n&\\le\n    \\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\sqrt{\\ell!} \\left(\\int_{\\bbR} \\left| f^{(\\tau-\\ell)}(x)\\right|^2\\rho(x)\\mathrm{d} x\\right)^{1/2}\\left(\\int_{\\bbR} \\left|H_{\\ell}(x) \\right|^2 \\rho(x) \\mathrm{d} x\\right)^{1/2}\n    <\\infty,\n\\end{align*}\nwhile for $\\tau=\\alpha$ we have\n\\begin{align*}\n    \\| g^{(\\alpha)}(x) \\|_{L^2(\\bbR)}\n    &\\le\n    \\sum_{\\ell=0}^\\alpha \\binom{\\alpha}{\\ell} \\| f^{(\\alpha-\\ell)}(x)\\rho^{(\\ell)}(x) \\|_{L^2(\\bbR)}\n    \\\\\n    &=\n    \\sum_{\\ell=0}^\\alpha \\binom{\\alpha}{\\ell} \\left(\\int_{\\bbR} \\left|\\rho(x) f^{(\\alpha-\\ell)}(x)\\sqrt{\\ell!}(-1)^{\\ell} H_{\\ell}(x) \\right|^2 \\mathrm{d} x\\right)^{1/2}\n    \\\\\n    &\\le\n    \\sum_{\\ell=0}^\\alpha \\binom{\\alpha}{\\ell} \\left( \\sup_{t\\in\\bbR} \\left(H_{\\ell}^2(t)\\rho(t) \\right) \\int_{\\bbR} \\rho(x) \\bigl|f^{(\\alpha-\\ell)}(x)\\bigr|^2\\ell!    \\mathrm{d} x\\right)^{1/2}\n    <\\infty.\n\\end{align*}\nHence, $\\|g\\|^*_\\alpha\\leq C \\|f\\|_{\\alpha}$ holds.\n\nTo show $\\|g\\|_{\\alpha,\\mathrm{decay}}\\leq C \\|f\\|_{\\alpha}$, let $h_{\\tau}:=\\rho^{\\varepsilon-1} g^{(\\tau)}$ for $\\tau=0,\\ldots,\\alpha-1$. Then, we have\n\\begin{align*}\n    \\| h_{\\tau} \\|_{L^2(\\bbR)}\n    &\\le\n    \\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\|\\rho^{\\varepsilon-1}(x) f^{(\\tau-\\ell)}(x)\\rho^{(\\ell)}(x) \\|_{L^2(\\bbR)}\n    \\\\\n    &=\n    \\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\left(\\int_{\\bbR} \\left|\\rho^{\\varepsilon}(x) f^{(\\tau-\\ell)}(x)\\sqrt{\\ell!}(-1)^{\\ell} H_{\\ell}(x) \\right|^2 \\mathrm{d} x\\right)^{1/2}\n    \\\\\n    &\\le\n    \\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\left( \\sup_{t\\in\\bbR} \\left(H_{\\ell}^2(t)\\rho^{2\\varepsilon-1}(t) \\right) \\int_{\\bbR} \\rho(x) \\left|f^{(\\tau-\\ell)}(x)\\right|^2 \\ell!   \\mathrm{d} x\\right)^{1/2} \n    <\\infty\n\\end{align*}\nand \n\\begin{align*}\n\\|&h_{\\tau}'(x)\\|_{L^2(\\bbR)}\n\\le\n\\|(1-\\varepsilon)x g^{(\\tau)}(x)\\rho^{\\varepsilon-1} (x) \\|_{L^2(\\bbR)} + \\|g^{(\\tau+1)}(x)\\rho^{\\varepsilon-1}(x)\\|_{L^2(\\bbR)}\n\\\\\n&\n\\le\n\\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell}\\left( \\int_{\\bbR} \\left|(1-\\varepsilon)x\\rho^{\\varepsilon}(x) f^{(\\tau-\\ell)}(x)\\sqrt{\\ell!}(-1)^{\\ell} H_{\\ell}(x) \\right|^2 \\mathrm{d} x\\right)^{1/2}\n\\\\\n&\\quad+\n\\sum_{\\ell=0}^{\\tau+1} \\binom{\\tau+1}{\\ell}\\left(\\int_{\\bbR} \\left|\\rho^{\\varepsilon}(x) f^{({\\tau+1}-\\ell)}(x)\\sqrt{\\ell!}(-1)^{\\ell} H_{\\ell}(x) \\right|^2 \\mathrm{d} x\\right)^{1/2} \\\\\n&\\le\n\\sum_{\\ell=0}^\\tau \\binom{\\tau}{\\ell} \\left( \\sup_{t\\in \\bbR}\\left|\\rho^{2\\varepsilon-1}(t)(1-\\varepsilon)^2 t^2 H^2_{\\ell}(t) \\right| \\int_{\\bbR} |f^{(\\tau-\\ell)}(x)|^2\\rho(x)\\ell! \\mathrm{d} x\\right)^{1/2}\\\\\n&\\quad+\n\\sum_{\\ell=0}^{\\tau+1} \\binom{\\tau+1}{\\ell}\\left(\\sup_{t\\in \\bbR}\\left| \\rho^{2\\varepsilon-1}(t) H_{\\ell}^2(t) \\right|  \\int_{\\bbR} |f^{({\\tau+1}-\\ell)}(x)|^2 \\ell! \\rho(x)\\mathrm{d} x\\right)^{1/2}<\n\\infty.\n\\end{align*}\nThus, from the Sobolev inequality, e.g., \\cite[Theorem~8.8]{B2011}, for $\\tau=0,\\dots,\\alpha-1$ we have $\\|h_{\\tau}\\|_\\infty\\le C\\|h_{\\tau}\\|_{W^{1,2}(\\bbR)}<\\infty$. This completes the proof.\n\\end{proof}\nTheorem~\\ref{thm:trape-opt} is an application of Proposition~\\ref{prop:trape} to $g=f\\rho$ with $f\\in \\mathscr{H}_{\\alpha}$.  \nSimilarly, applying Corollary~\\ref{cor:alpha-free} to $g=f\\rho$ yields a trapezoidal rule whose construction is independent of $\\alpha$ with the optimal convergence rate up to a factor of $(\\gamma(n)\\ln n)^{ (\\alpha/2 + 1/4)}$ for functions in $\\mathscr{H}_{\\alpha}$. \nSince the argument is straightforward from \nCorollary~\\ref{cor:alpha-free} and \nTheorem~\\ref{thm:trape-opt}, we omit the details.\n\n\\subsection*{Details of Figure~\\ref{fig:GH-sub}}\nNow we are ready to discuss the details of Figure~\\ref{fig:GH-sub} in Section~\\ref{sec:intro}. \nThe trapezoidal rule used there is $Q^*_{n,T}$ as in \\eqref{eq:def-trap} with $T=\\sqrt{\\frac{2}{1-\\varepsilon}\\alpha\\ln (n)}$ and $\\varepsilon=0.51$. \nHere, we chose  $\\alpha=p$, since $f(x)=|x|^p$ is in $\\mathscr{H}_p$ but not in $\\mathscr{H}_{p+1}$.\nThe number of points $n$ is chosen to be odd for the trapezoidal rule and even for Gauss--Hermite quadrature, so that both quadrature rules do not evaluate at the origin $x=0$ where the integrand is not smooth. \n\nThe rate around $\\mathcal{O}(n^{-p/2-0.5})$ we observe for Gauss--Hermite quadrature is consistent with the matching bounds of the of order $n^{-\\alpha/2}=n^{-p/2}$ in the sense of worst-case error, since $f(x)=|x|^p$ is a specific element from $\\mathscr{H}_p$; \nthe rate around $\\mathcal{O}(n^{-p-0.8})$ we observe for the trapezoidal rule also supports our results,  according to which we expect to see at least $\\mathcal{O}(n^{-p})$ for any function in $\\mathscr{H}_p$.\n\n\\section{Conclusions}\\label{sec:conc}\nIn this paper, we proved the sub-optimality of Gauss--Hermite quadrature and the optimality of the trapezoidal rule for functions with finite smoothness, in the sense of worst-case error.\nThe lower bound presented for Gauss--Hermite quadrature is sharp, and the upper bound presented for the trapezoidal rule is also sharp, up to a logarithmic factor.\n\nTo establish the lower bound for Gauss--Hermite rule, we constructed a sequence of fooling functions. \nThis strategy also demonstrated that what causes this lower bound is the placement of quadrature points, and thus tuning the quadrature weights does not improve the bound. \n\nA key for showing the optimality of the trapezoidal rule was the auxiliary periodic function in Lemma~\\ref{lem:trap-error-inside}. \nThe function used there is in fact an orthogonal projection in a suitable sense;  \nfor details, we refer to \\cite{NS2021}. \nNeedless to say, upon the domain truncation $[-T,T]$, other quadrature rules on the finite interval, such as Clenshaw--Curtis or Gauss--Legendre quadratures, can also be used. \nFor these quadrature rules, analogous upper bounds should be able to be derived, without the necessity of introducing the aforementioned periodic function. \nSince these quadratures are arguably more complicated to use than the trapezoidal rule, and the error analysis should be less involved, we had left them out from the scope of this paper.\n\nOur results suggest that the truncated trapezoidal rule may be also promising for high-dimensional problems. \nOne generalisation of the trapezoidal rule to the multidimensional setting is the lattice rule. \nNuyens and Suzuki \\cite{NS2021} studied this method for the integration problem on $\\mathbb{R}^d$ with respect to the Lebesgue measure. \nTo verify if it works well for the Gaussian measure in a high-dimensional setting is kept for future works.\n\nAnother generalisation to high-dimensional settings is by the Smolyak-type algorithms. \nAs mentioned in Section~\\ref{sec:intro}, this type of methods based on Gauss--Hermite points is widely used. \nIn light of the results presented in this paper, especially when the target function is expected to have limited smoothness, the trapezoidal rule may be a better choice.\nInvestigating these speculations is also kept for future works.\n\n\\section*{Acknowledgments}\nPart of this work was carried out when Yoshihito Kazashi was working at CSQI, Institute of Mathematics, \\'Ecole Polytechnique F\\'ed\\'erale de Lausanne, Switzerland. We thank Dirk Nuyens and Ken’ichiro Tanaka for their valuable comments.\n\n\\bibliographystyle{siamplain}\n", "meta": {"timestamp": "2023-01-16T02:11:10", "yymm": "2202", "arxiv_id": "2202.11420", "language": "en", "url": "https://arxiv.org/abs/2202.11420", "abstract": "The sub-optimality of Gauss--Hermite quadrature and the optimality of the trapezoidal rule are proved in the weighted Sobolev spaces of square integrable functions of order $\\alpha$, where the optimality is in the sense of worst-case error. For Gauss--Hermite quadrature, we obtain matching lower and upper bounds, which turn out to be merely of the order $n^{-\\alpha/2}$ with $n$ function evaluations, although the optimal rate for the best possible linear quadrature is known to be $n^{-\\alpha}$. Our proof of the lower bound exploits the structure of the Gauss--Hermite nodes; the bound is independent of the quadrature weights, and changing the Gauss--Hermite weights cannot improve the rate $n^{-\\alpha/2}$. In contrast, we show that a suitably truncated trapezoidal rule achieves the optimal rate up to a logarithmic factor.", "subjects": "Numerical Analysis (math.NA)", "title": "Sub-optimality of Gauss--Hermite quadrature and optimality of the trapezoidal rule for functions with finite smoothness", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9780517488441416, "lm_q2_score": 0.8670357512127872, "lm_q1q2_score": 0.8480058327840606}}
{"url": "https://arxiv.org/abs/1207.6681", "title": "Box-counting fractal strings, zeta functions, and equivalent forms of Minkowski dimension", "abstract": "We discuss a number of techniques for determining the Minkowski dimension of bounded subsets of some Euclidean space of any dimension, including: the box-counting dimension and equivalent definitions based on various box-counting functions; the similarity dimension via the Moran equation (at least in the case of self-similar sets); the order of the (box-)counting function; the classic result on compact subsets of the real line due to Besicovitch and Taylor, as adapted to the theory of fractal strings; and the abscissae of convergence of new classes of zeta functions. Specifically, we define box-counting zeta functions of infinite bounded subsets of Euclidean space and discuss results pertaining to distance and tube zeta functions. Appealing to an analysis of these zeta functions allows for the development of theories of complex dimensions for bounded sets in Euclidean space, extending techniques and results regarding (ordinary) fractal strings obtained by the first author and van Frankenhuijsen.", "text": "\\section{Introduction}\n\\label{sec:Introduction}\n\nMotivated by the theory of complex dimensions of fractals strings (the main theme of \\cite{LapvF6}), we introduce box-counting fractal strings and box-counting zeta functions which, along with the distance and tube zeta functions of \\cite{LapRaZu}, provide possible foundations for the pursuit of theories of complex dimensions for arbitrary bounded sets in Euclidean space of any dimension. We also summarize a variety of well-known techniques for determining the box-counting dimension, or equivalently the Minkowski dimension, of such sets. Thus, while new results are presented in this paper, it is partially expository and also partially tutorial.\n\nOur main result establishes line (iv) of the following theorem. (See also Theorem \\ref{thm:MainResult} below, along with the relevant definitions provided in this paper.) The other lines have been established elsewhere in the literature, as cited accordingly throughout the paper.\n\n\\begin{theorem}\n\\label{thm:Summary}\nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$ \\emph{(}equipped with the usual metric\\emph{)}. Then the following quantities are equal\\emph{:}\\footnote{The fact that $A$ is infinite is only required for part (iv) of the theorem; see Remark \\ref{rmk:FiniteBoxCountingFractalString}.}\n\\begin{enumerate}\n\\setlength{\\itemsep}{0in}\n\\item the \\emph{upper box-counting dimension} of $A$\\emph{;}\n\\item the \\emph{upper Minkowski dimension} of $A$\\emph{;}\n\\item the \\emph{asymptotic order of growth of the counting function} of the box-counting fractal string $\\mathcal{L}_B$\\emph{;}\n\\item the \\emph{abscissa of convergence} of the box-counting zeta function $\\zeta_B$\\emph{;}\n\\item the \\emph{abscissa of convergence} of the distance zeta function $\\zeta_d$.\n\\end{enumerate}\n\\end{theorem}\n\nA summary of the remaining sections of this paper is as follows:\n \nIn Section \\ref{sec:ClassicNotionsOfDimension}, we discuss classical notions of dimension such as \\emph{similarity dimension} (or \\emph{exponent}), \\emph{box-counting dimension}, and \\emph{Minkowski dimension} as well as their properties. (See \\cite{BesTa,Falc,HaPo,Hut,LapLuvF1,LapLuvF2,LapPe3,LapPe2,LapPeWin,LapPo1,LapRaZu,\nLapvF6,MeSa,Mor,Tr2,Tri,Zu,ZuZup}.)\n\nIn Section \\ref{sec:FractalStringsAndZetaFunctions}, we summarize but a few of the interesting results on fractal strings and counting functions regarding, among other things, geometric zeta functions, complex dimensions, the order of a counting function, and connections with Minkowski measurability. (See \\cite{BesTa,LapPo1,LapvF6,Lev,Tri}.) The material in Sections \\ref{sec:ClassicNotionsOfDimension} and \\ref{sec:FractalStringsAndZetaFunctions} motivates the results presented in Sections \\ref{sec:BoxCountingFractalStringsAndZetaFunctions} and \\ref{sec:DistanceAndTubeZetaFunctions}.\n\nIn Section \\ref{sec:BoxCountingFractalStringsAndZetaFunctions}, we introduce \\emph{box-counting fractal strings} and \\emph{box-counting zeta functions} and, in particular, we show that the abscissa of convergence of the box-counting zeta function of a bounded infinite set is the upper box-counting dimension of the set. These topics are the focus of \\cite{LapRoZu}.\n\nIn Section \\ref{sec:DistanceAndTubeZetaFunctions}, we share recent results from \\cite{LapRaZu} on \\emph{distance, tube} and \\emph{relative zeta functions}, including connections between the corresponding complex dimensions and Minkowski content and measurability.\n\nIn Section \\ref{sec:SummaryOfResultsAndOpenProblems}, Theorem \\ref{thm:Summary} is restated in Theorem \\ref{thm:MainResult} using notation and terminology discussed throughout the paper. We also propose several open problems for future work in this area.\n\n\n\\section{Classic notions of dimension}\n\\label{sec:ClassicNotionsOfDimension}\n\nWe begin with a brief discussion of a classic method for constructing self-similar fractals and a famous fractal, the Cantor set $C$. (See \\cite{Falc,Hut}.)\n\n\\begin{definition}\n\\label{def:IFSAndAttractor}\nLet $N$ be an integer such that $N \\geq 2$. An \\emph{iterated function system} (IFS) $\\mathbf{\\Phi}=\\{\\Phi_j\\}_{j=1}^N$ is a finite family of contractions on a complete metric space $(X,d_X)$. Thus, for all $x,y \\in X$ and each $j=1,\\ldots,N$ we have \n\\begin{align}\nd_X(\\Phi_j(x),\\Phi_j(y))\t&\\leq r_jd_X(x,y),\n\\end{align}\nwhere $0<r_j<1$ is the \\emph{scaling ratio} (or Lipschitz constant) of $\\Phi_j$ for each $j=1,\\ldots,N$. \n\nThe \\emph{attractor} of $\\mathbf{\\Phi}$ is the nonempty compact set $F \\subset X$ defined as the unique fixed point of the contraction mapping \n\\begin{align}\n\\label{eqn:IFSAsMap}\n\\displaystyle\n\\mathbf{\\Phi}(\\cdot)\t&:= \\bigcup_{j=1}^N \\Phi_j(\\cdot)\n\\end{align}\non the space of compact subsets of $X$ equipped with the Hausdorff metric. That is, $F=\\mathbf{\\Phi}(F)$. If \n\\begin{align}\n\\label{eqn:IFSSelfSimilar}\nd_X(\\Phi_j(x),\\Phi_j(y)) &= r_jd_X(x,y)\n\\end{align}\nfor each $j=1,\\ldots,N$ (i.e., if the contraction maps $\\Phi_j$ are similarities with scaling ratios $r_j$), then the attractor $F$ is the \\emph{self-similar set} associated with $\\mathbf{\\Phi}$.\n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:AttentionOnEuclideanSpacesAndOSC}\nWe focus our attention on Euclidean spaces of the form $X=\\mathbb{R}^m$, where $m$ is a positive integer and $d_X=d_m$ is the classic $m$-dimensional Euclidean distance. We  denote $d_m(x,y)$ by $|x-y|$. Furthermore, we consider only iterated functions systems which satisfy the \\emph{open set condition} (see \\cite{Falc,Hut}). Recall that an IFS $\\mathbf{\\Phi}$ satisfies the open set condition if there is a nonempty open set $V \\subset \\mathbb{R}^m$ for which $\\Phi_j(V)\\subset V$ for each $j$ and the images $\\Phi_j(V)$ are pairwise disjoint.\n\\end{remark}\n\n\n\\begin{figure}\n\\includegraphics[scale=.5]{cantorstringharp.eps}\n\\caption{The classic ``middle-third removal'' construction of the Cantor set $C$ is depicted on the left. The Cantor string $\\mathcal{L}_{CS}$ is the nonincreasing sequence comprising the lengths of the removed intervals which are depicted on the right as a fractal harp.}\n\\label{fig:CantorConstructionAndHarp}\n\\end{figure}\n\n\\begin{example}[The dimension of the Cantor set]\nThe Cantor set $C$ can be constructed in various ways. For instance, we have the classic ``middle-third removal'' construction of $C$ as depicted in Figure 1. \nA more elegant construction shows $C$ to be the unique nonempty attractor of the iterated function system $\\mathbf{\\Phi}_C$ on $[0,1]$ given by the two contracting similarities $\\varphi_1(x)=x/3$ and $\\varphi_2(x)=x/3+2/3$. The box-counting dimension of $C$ is $\\log_{3}2$, a fact which can be established with any of the myriad of formulas presented in this paper. Notably, $\\log_{3}2$ is equivalently found to be: the \\emph{order of the geometric counting function} (see Remark \\ref{rmk:DimensionEqualsOrder}) of the \\emph{box-counting fractal string} $\\mathcal{L}_B$ of $C$ (which is related but not equal to the Cantor string $\\mathcal{L}_{CS}$, see Definition \\ref{def:BoxCountingFractalString}, Equation \\eqref{eqn:CantorStringLengths}, and \\cite[Ch.1]{LapvF6}); the \\emph{abscissa of convergence} of either the \\emph{geometric zeta function} of $\\mathcal{L}_{CS}$ (Definition \\ref{def:GeometricZetaFunction}), the \\emph{box-counting zeta function} of $C$ (Definition \\ref{def:BoxCountingZetaFunction}), the \\emph{distance zeta function} of $C$ (Definition \\ref{def:DistanceZetaFunction}), or the \\emph{tube zeta function} of $C$ (Equation \\eqref{tube_zeta} in Section \\ref{sec:TubeZetaFunction}); or else the unique real-valued solution of the corresponding Moran equation (cf. Equation \\eqref{eqn:Moran}):  $2\\cdot3^{-s}=1$.\n\\end{example}\n\n\n\\subsection{Similarity dimension}\n\\label{sec:SimilarityDimension}\n\nThe first notion of `dimension' we consider is the similarity dimension (or `similarity exponent') of a self-similar set.\n\n\\begin{definition}\n\\label{def:SimilarityDimension}\nLet $\\mathbf{\\Phi}$ be an iterated function system that satisfies the open set condition and is generated by similarities with scaling ratios $\\{r_j\\}_{j=1}^N$, with $N \\geq 2$. Then the \\emph{similarity dimension} of the attractor of $\\mathbf{\\Phi}$ (that is, of the self-similar set associated with $\\mathbf{\\Phi}$) is the unique real solution $D_\\mathbf{\\Phi}$ of the equation  \n\\begin{align}\n\\label{eqn:Moran}\n\\sum_{j=1}^N r_j^\\sigma\t&= 1, \\quad \\sigma \\in \\mathbb{R}.\n\\end{align}\n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:MoransTheorem}\nEquation \\eqref{eqn:Moran} is known as Moran's equation. Moran's Theorem is a well-known result which states that the similarity dimension $D_\\mathbf{\\Phi}$ is equal to the box-counting (and Hausdorff) dimension of the self-similar attractor of $\\mathbf{\\Phi}$.\\footnote{Moran's original result in \\cite{Mor} was established in $\\mathbb{R}$ (i.e., for $m=1$) but is valid for $m\\geq 1$; cf. \\cite{Falc,Hut}.} In fact, $D_\\mathbf{\\Phi}$ is positive, a fact that can be verified directly from Equation \\eqref{eqn:Moran}. For details regarding iterated functions systems, the open set condition, and Moran's Theorem, see \\cite[Ch.9]{Falc} as well as \\cite{Hut} and \\cite{Mor}.\n\\end{remark}\n\n\n\\subsection{Box-counting dimension}\n\\label{sec:BoxCountingDimension}\n\nIn this section we discuss the central notion of box-counting dimension and some of its properties.\n\n\\begin{definition}\n\\label{def:BoxCountingFunction}\nLet $A$ be a subset of $\\mathbb{R}^m$. The \\emph{box-counting function} of $A$ is the function $N_B(A,\\cdot): (0,\\infty) \\rightarrow \\mathbb{N} \\cup \\{0\\}$, where (for a given $x>0$) $N_B(A,x)$ denotes the maximum number of disjoint closed balls $B(a,x^{-1})$ with centers $a \\in A$ of radius $x^{-1}$.\n\\end{definition}\n\n\\begin{definition}\n\\label{def:BoxCountingDimension}\nFor a set $A \\subset \\mathbb{R}^m$, the \\emph{lower} and \\emph{upper box-counting dimensions} of $A$, denoted $\\underline{\\dim}_{B}A$ and $\\overline{\\dim}_{B}A$, respectively, are given by\n\\begin{equation}\n\\begin{aligned}\n\\underline{\\dim}_{B}A&:= \\liminf_{x\\to\\infty}\\frac{\\log N_B(A,x)}{\\log x}, \\label{eqn:UpperBoxDimension}\\\\\n\\overline{\\dim}_{B}A&:= \\limsup_{x\\to\\infty}\\frac{\\log N_B(A,x)}{\\log x}. \n\\end{aligned}\n\\end{equation}\n\n\n\nWhen $\\underline{\\dim}_{B}A=\\overline{\\dim}_{B}A$, the following limit exists and is called the \\emph{box-counting dimension} of $A$, denoted $\\dim_{B}A$:\n\\begin{align*}\n\\dim_{B}A\t&:= \\lim_{x\\to\\infty}\\frac{\\log N_B(A,x)}{\\log x}.\n\\end{align*}\n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:EquivalentUpperBoxDimension}\nThe upper and lower box-counting dimensions are independent of the ambient dimension $m$. In most applications the set $A$ is such that $N_B(A,x) \\asymp x^d$ as $x\\to\\infty$, for some constant $d \\in [0,m]$ (the relation $\\asymp$ is explained at the end of Notation \\ref{not:DistanceVolumeAndBigO} below). It is easy to see that then, $\\dim_{B}A=d$. Also, the following equivalent form of the upper box-counting dimension will prove to be useful in Section \\ref{sec:BoxCountingFractalStringsAndZetaFunctions} (see Notation \\ref{not:DistanceVolumeAndBigO} below for an explanation of `Big-$O$' notation: $f(x)=O(g(x))$ as $x \\rightarrow \\infty$).\n\\begin{align*}\n\\overline{\\dim}_{B}A\t&=\\inf\\{\\alpha \\geq 0 : N_B(A,x)=O(x^{\\alpha}) \\textnormal{ as } x \\rightarrow \\infty\\}.\n\\end{align*}\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:VariousCountingFunctions} \nThere are many equivalent definitions of the box-counting dimension (see \\cite[Ch.3]{Falc}). For instance, the box-counting function $N_B(A,x)$ given in Definition \\ref{def:BoxCountingFunction} may be replaced by:\n\\begin{enumerate}\n\\setlength{\\itemsep}{0in}\n\\item the minimum number of sets of diameter at most $x^{-1}$ required to cover $A$;\n\\item the minimum number of closed balls of radius $x^{-1}$ required to cover $A$;\n\\item the minimum number of closed cubes with side length $x^{-1}$ required to cover $A$; or\n\\item the number of $x^{-1}$-mesh cubes that intersect $A$.\n\\end{enumerate}\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:varepsilonVSx}\nOne may also define the box-counting function in terms of $\\varepsilon>0$, where $\\varepsilon=x^{-1}$ plays the role of the scale under consideration.\\footnote{Indeed, note that given $\\varepsilon>0$, $N_B(A,\\varepsilon^{-1})$ is the maximum number of disjoint balls $B(a,\\varepsilon)$ with center $a \\in A$ and radius $\\varepsilon$ (or, \\emph{mutatis mutandis}, $\\varepsilon$ denotes any of the counterparts to the notion of radius given in (i)--(iv) of Remark \\ref{rmk:VariousCountingFunctions}).} Although this may be a more natural way to describe a box-counting function, the results relating box-counting functions and \\emph{geometric counting functions} (see Definition \\ref{def:GeometricCountingFunction}) presented in Section \\ref{sec:BoxCountingFractalStringsAndZetaFunctions} are stated and analyzed in terms of $x>0$. Moreover, this convention is used throughout \\cite{LapvF6} and that text will be vital to the development of further material based on the results presented in this paper.\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:PropertiesBoxCountingDimension}\nIf the box-counting function $N_B(A,x)$ is given as in Definition \\ref{def:BoxCountingFunction} or one of the alternatives in Remark \\ref{rmk:VariousCountingFunctions}, then the upper and lower box-counting dimensions have the following properties (cf. \\cite[Ch.~3]{Falc} and \\cite{Fdrr}):\n\\begin{enumerate}\n\\setlength{\\itemsep}{0in}\n\\item Let $V$ be a bounded $n$-dimensional submanifold of $\\mathbb{R}^m$ which is \\emph{rectifiable} in the sense that $V \\subset f(\\mathbb{R}^n)$, where $f:\\mathbb{R}^n\\to\\mathbb{R}^m$ is a Lipschitz function. Then $\\dim_{B}V=n$.\n\\item Both $\\overline{\\dim}_B$ and $\\underline{\\dim}_B$ are monotonic. That is, if $A_1 \\subset A_2 \\subset \\mathbb{R}^m$, then\n\t\\begin{align*}\n\n\t\\overline{\\dim}_{B}A_1 &\\leq \\overline{\\dim}_{B}A_2, \\quad \t\t\t \\underline{\\dim}_{B}A_1 \\leq \\underline{\\dim}_{B}A_2.\n\t\\end{align*}\n\n\\item Let $\\overline{A}$ denote the closure of $A$ (i.e., the smallest closed subset of $\\mathbb{R}^m$ which contains $A$). Then \n\t\\begin{align*}\n\n\t\\overline{\\dim}_{B}\\overline{A}\t&= \\overline{\\dim}_{B}A, \\quad\n\t\\underline{\\dim}_{B}\\overline{A} = \\underline{\\dim}_{B}A.\n\t\\end{align*}\n\\item  For any two sets $A_1,A_2 \\subset \\mathbb{R}^m$,\n\t\\begin{align*}\n\n\t\\overline{\\dim}_{B}(A_1 \\cup A_2)&= \\max\\left\\{\\overline{\\dim}_{B}A_1,\\overline{\\dim}_{B}A_2\\right\\}.  \n\t\\end{align*} That is, $\\overline{\\dim}_B$ is finitely stable. On the other hand, $\\underline{\\dim}_B$ is not finitely stable. \n\\item Neither $\\overline{\\dim}_B$ nor $\\underline{\\dim}_B$ is countably stable. That is, neither $\\overline{\\dim}_B$ nor $\\underline{\\dim}_B$ satisfies the analogue of property (iv) for a countable collection of subsets of $\\mathbb{R}^m$. \t\n\\end{enumerate}\n\\end{remark}\n\nConcerning the loss of finite stability of the lower box dimension mentioned in\n(iv) of Remark \\ref{rmk:PropertiesBoxCountingDimension}, it is possible to construct two bounded sets $A$ and $B$ in $\\mathbb{R}^m$ such that their lower box-counting dimensions are both equal to zero, while the box-counting dimension of their union is equal to $m$; see \\cite[Theorem 1.4]{Zu}.\n\nA simple way to see why property (v) of Remark \\ref{rmk:PropertiesBoxCountingDimension} is satisfied for the upper box-counting dimension is to consider the countable set $A=\\{1,1/2,1/3,\\ldots\\}$ and note that $\\overline{\\dim}_{B}A=1/2$ whereas $\\overline{\\dim}_{B}\\{1/j\\}=0$ for each positive integer $j$.\n\nThe following proposition shows that one need only consider certain discrete sequences of scales which tend to zero in order to determine the box-counting dimension of a set.\n\n\\begin{proposition}\n\\label{prop:GeometricSequenceSufficient}\nLet $\\lambda>1$ and $A \\subset \\mathbb{R}^m$. Then\n\\begin{align*}\n\\underline{\\dim}_{B}A\t&=\\liminf_{k \\rightarrow \\infty}\\frac{\\log N_B(A,\\lambda^{k})}{\\log\\lambda^{k}}, \\\\\n\\overline{\\dim}_{B}A\t&=\\limsup_{k \\rightarrow \\infty}\\frac{\\log N_B(A,\\lambda^{k})}{\\log\\lambda^{k}}.\n\\end{align*}\n\\end{proposition}\n\n\\begin{proof}\nIf $\\lambda^{k}< x \\leq\\lambda^{k+1}$, then\n\\begin{align*}\n\\frac{\\log N_B(A,x)}{\\log x} &\\leq  \n\\frac{\\log N_B(A,\\lambda^{k+1})}{\\log\\lambda^{k}} = \n\\frac{\\log N_B(A,\\lambda^{k+1})}{\\log\\lambda^{k+1}-\\log\\lambda}.\n\\end{align*}\nTherefore, \n\\begin{align*}\n\\limsup_{x\\to\\infty}\\frac{\\log N_B(A, x)}{\\log  x}\t& \\leq \\limsup_{k \\rightarrow \\infty}\\frac{\\log N_B(A,\\lambda^{k})}{\\log\\lambda^{k}}.\n\\end{align*}\nThe opposite inequality clearly holds and the case for the lower limits follows \\emph{mutatis mutandis}.\n\\end{proof}\n\n\\begin{example}[Box-counting dimension of the Cantor set]\n\\label{eg:BoxCountingDimensionOfCantorSet}\nLet $C$ be the Cantor set and $n \\in \\mathbb{N}$. Also, let $N_B(A,3^n)$ denote the minimum number of disjoint closed intervals with length $3^{-n}$ required to cover $C$. Then\n$N_B(A,3^n)=2^n$, so by Proposition \\ref{prop:GeometricSequenceSufficient} we have\n\\begin{align}\n\\label{eqn:BoxCountingDimensionOfCantorSet}\n\\dim_{B}C\t&= \\lim_{n\\rightarrow\\infty} \\frac{\\log{2^n}}{\\log{3^n}}=\\log_{3}2.\n\\end{align}\n\\end{example}\n\nIn the next section, we discuss the Minkowski dimension, which is well known to be equivalent to the box-counting dimension.\n\n\n\\subsection{Minkowski dimension}\n\\label{sec:MinkowskiDimension}\n\nMinkowski content and Minkowski dimension require a specific notion of volume and can be stated concisely with the following notation.\n\n\\begin{notation}[Distance, volume, and Big-$O$]\n\\label{not:DistanceVolumeAndBigO}\nLet $\\varepsilon >0$ and $A \\subset \\mathbb{R}^m$. Let $d(x,A)$ denote the distance between a point $x\\in\\mathbb{R}^m$ and the set $A$ given by\n\\begin{align*}\nd(x,A)\t&:= \\inf\\{|x-u|_m : u \\in A\\},\n\\end{align*}\nwhere $|\\cdot|_m$ denotes the $m$-dimensional Euclidean norm. The \\emph{$\\varepsilon$-neighborhood} of $A$, denoted $A_\\varepsilon$, is the set of points in $\\mathbb{R}^m$ which are within $\\varepsilon$ of $A$. Specifically,\n\\begin{align*}\nA_\\varepsilon\t&= \\{x \\in \\mathbb{R}^m: d(x,A) < \\varepsilon\\}.\n\\end{align*}\n\nIn the sequel, we fix the set $A$ and are concerned with the $m$-dimensional Lebesgue measure (denoted $\\textnormal{vol}_m$) of its $\\varepsilon$-neighborhood $A_\\varepsilon$ for a given $\\varepsilon>0$. Recall, for completeness,\\footnote{The book \\cite{Cohn} is a good general reference on elementary (as well as more advanced) measure theory.} that the $m$-dimensional Lebesgue measure of a (measurable) set $A \\subset \\mathbb{R}^m$ is given by \n\\begin{align*}\n\\textnormal{vol}_m(A)\t&:=\\inf\\left\\{\\sum_{n=1}^\\infty \\prod_{j=1}^m(b_{n,j}-a_{n,j}) : A \\subset \\bigcup_{n=1}^\\infty \\left(\\prod_{j=1}^m[a_{n,j},b_{n,j}]\\right) \\right\\}.\n\\end{align*}\n\nIn the case of an \\emph{ordinary fractal string} $\\Omega \\subset \\mathbb{R}$ (see the latter part of Definition \\ref{def:FractalString}), we are interested in the 1-dimensional volume (i.e., length) of the  \\emph{inner} $\\varepsilon$-neighborhood of the boundary $\\partial\\Omega$. Specifically, given an ordinary fractal string $\\Omega$ and $\\varepsilon>0$, the volume $V_{\\textnormal{inner}}(\\varepsilon)$ of the inner $\\varepsilon$-neighborhood of $\\partial\\Omega$ is defined by \n\\begin{align}\n\\label{eqn:VolumeInnerNeighborhood}\nV_{\\textnormal{inner}}(\\varepsilon)\t&:=\\textnormal{vol}_1\\{x \\in \\Omega : d(x,\\partial\\Omega) <\\varepsilon\\}.\n\\end{align}\n\nFor two functions $f$ and $g$, with $g$ nonnegative, we write $f(x)=O(g(x))$ as $x\\to\\infty$ if there exists a positive real number $c$ such that for all sufficiently large $x$, $|f(x)|\\leq cg(x)$. More generally, if there exists $C$ such that $|f(x)|\\leq Cg(x)$ for all $x$ sufficiently close to some value $a \\in \\mathbb{R}\\cup \\{\\pm\\infty\\}$, then we write $f(x)=O(g(x))$ as $x\\to a$. If both $f(x)=O(g(x))$ and $g(x)=O(|f(x)|)$ as $x\\to a$, we write $f(x)\\asymp g(x)$ as $x\\to a$. Moreover, if $\\lim_{x\\to a} f(x)/g(x)=1$,\\footnote{Or, more generally, if $f(x)=g(x)(1+o(1))$ as $x\\to a$, where $o(1)$ stands for a function tending to zero as $x\\to a$.} then we write $f(x) \\sim g(x)$ as $x\\to a$. Analogous notation will be used for infinite sequences.\n\\end{notation}\n\n\\begin{definition}[Minkowski content]\n\\label{def:MinkowskiContent0}\nLet $r$ be a given nonnegative real number. The \\emph{upper} and \\emph{lower $r$-dimensional Minkowski contents} of a bounded set $A\\subset \\mathbb{R}^m$ are respectively given by\n\\begin{align*}\n\\mathscr{M}^{*r}(A) &:= \\limsup_{\\varepsilon\\rightarrow 0^+}\\frac{\\textnormal{vol}_m(A_\\varepsilon)}{\\varepsilon^{m-r}},\\\\\n\\mathscr{M}_*^{r}(A) &:= \\liminf_{\\varepsilon\\rightarrow 0^+}\\frac{\\textnormal{vol}_m(A_\\varepsilon)}{\\varepsilon^{m-r}}.\n\\end{align*}\n\\end{definition}\n\nIt is easy to see that if $\\mathscr{M}^{*r}(A)<\\infty$, then $\\mathscr{M}^{*s}(A)=0$ for each $s>r$. Furthermore, since $A$ is bounded, then clearly $\\mathscr{M}^{*r}(A)=0$ for $r>m$. On the other hand, if $\\mathscr{M}^{*r}(A)>0$, then $\\mathscr{M}^{*s}(A)=\\infty$ for each $s<r$. Therefore, there exists a unique point in $[0,m]$ at which the function $r\\mapsto \\mathscr{M}^{*r}(A)$ jumps from the value of $\\infty$ to zero. This unique point is called the \\emph{upper Minkowski dimension} of $A$. The \\emph{lower Minkowski dimension} of $A$ is defined analogously by using the lower $r$-dimensional Minkowski content.\n\n\\begin{definition}[Minkowski dimension]\n\\label{def:MinkowskiDimension0}\nThe \\emph{upper} and \\emph{lower Minkowski dimensions} of a bounded set $A$  are defined respectively by\n\\begin{equation}\n\\begin{aligned}\n\\overline{\\dim}_{M}A \t&:=\\inf\\{r\\ge0:\\mathscr{M}^{r*}(A)=0\\}=\\sup\\{r\\ge0:\\mathscr{M}^{r*}(A)=\\infty\\}, \\label{eqn:UpperMinkowskiDimension}\\\\\n\\underline{\\dim}_{M}A \t&:= \\inf\\{r\\ge0:\\mathscr{M}_*^{r}(A)=0\\}=\\sup\\{r\\ge0:\\mathscr{M}_*^{r}(A)=\\infty\\}.\n\\end{aligned}\n\\end{equation}\nWhen $\\overline{\\dim}_{M}A=\\underline{\\dim}_{M}A$, the common value is called the \\emph{Minkowski dimension} of $A$, denoted by $\\dim_{M}A$.\n\\end{definition}\n\nWhen we write $\\dim_{M}A$, we implicitly assume that the Minkowski dimension of $A$  exists. In most applications we have that $\\textnormal{vol}_m(A_\\varepsilon)\\asymp\\varepsilon^{\\alpha}$ as $\\varepsilon\\to0^+$, where $\\alpha$ is a number in $[0,m]$. Then $\\dim_{M}A$ exists and is equal to $m-\\alpha$ (in light of Definitions \\ref{def:MinkowskiContent0} and \\ref{def:MinkowskiDimension0}). Note that here \n\\begin{equation*}\n\\alpha=\\lim_{\\varepsilon\\to0^+}\\frac{\\log\\textnormal{vol}_m(A_\\varepsilon)}{\\log\\varepsilon},\n\\end{equation*}\nand hence, \n\\begin{equation*}\n\\dim_{M}A=m-\\lim_{\\varepsilon\\to0^+}\\frac{\\log\\textnormal{vol}_m(A_\\varepsilon)}{\\log\\varepsilon}.\n\\end{equation*}\nIt is not difficult to show that the following more general result holds.\n\n\\begin{proposition}\n\\label{prop:MinkowskiDimension}\nThe upper and lower Minkowski dimensions of a bounded set $A\\subset \\mathbb{R}^m$ are respectively given by\n\\begin{align*}\n\\overline{\\dim}_{M}A &= m-\\liminf_{\\varepsilon \\rightarrow 0^+}\\frac{\\log\\textnormal{vol}_m(A_\\varepsilon)}{\\log\\varepsilon},\\\\\n\\underline{\\dim}_{M}A &= m-\\limsup_{\\varepsilon \\rightarrow 0^+}\\frac{\\log\\textnormal{vol}_m(A_\\varepsilon)}{\\log\\varepsilon}.\n\\end{align*}\n\\end{proposition}\n\n\\begin{remark}\n\\label{rmk:EquivalentUpperMinkoskiDimension}\nThe upper and lower Minkowski dimensions are, of course, independent of the ambient dimension $m$. The upper Minkowski dimension is equivalently given by\n\\begin{align*}\n\\overline{\\dim}_{M}A\t&=\\inf\\{\\alpha \\geq 0 : \\textnormal{vol}_m(A_\\varepsilon)=O(\\varepsilon^{m-\\alpha}) \\textnormal{ as } \\varepsilon \\rightarrow 0^+\\}.\n\\end{align*}\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:Tr2}\nIt is interesting that there exists a bounded set $A$ in $\\mathbb{R}^m$ such that the upper and lower box dimension are different (see, e.g., \\cite[p.\\ 122]{Tr2}), and even such that $\\overline{\\dim}_{M}A=m$ and $\\underline{\\dim}_{M}A=0$ (see \\cite[Theorem 1.2]{Zu}).\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:HaPo}\nThe upper Minkowski dimension of $A$ is important in the study of the Lebesgue integrability of the distance function $d(x,A)^{-\\gamma}$ in an $\\varepsilon$-neighborhood of $A$, where $\\varepsilon$ is a fixed positive number:\n\\begin{equation}\n\\mbox{If $\\gamma<m-\\overline\\dim_{M}A$, then $\\int_{A_{\\varepsilon}}d(x,A)^{-\\gamma}dx<\\infty$.}\n\\end{equation} \nThis nice result is due to Harvey and Polking, and is implicitly stated in \\cite{HaPo}; see also \\cite{Zu} for related results and references. This fact enabled the first and third authors, along with G.~Radunovi$\\acute{\\textnormal{c}}$, to determine the abscissa of convergence of the so-called distance zeta function of $A$; see Definition~\\ref{def:DistanceZetaFunction} below along with Theorem \\ref{thm:DistanceZetaFunctionUpperBoxDimension} and \\cite{LapRaZu} for details.\n\\end{remark}\n\n\\begin{definition}[Minkowski measurability]\n\\label{def:MinkowskiMeasurability}\nLet $A \\subset \\mathbb{R}^m$ be such that $D_M=\\dim_{M}A$ exists. The \\emph{upper} and \\emph{lower Minkowski content} of $A$ are respectively defined as its $D_M$-dimensional upper \nand lower Minkowski contents, that is,\n\\begin{align*}\n\\mathscr{M}^* &:=\\mathscr{M}^{*D_M}(A)=\\limsup_{\\varepsilon\\to 0^+} \\frac{\\textnormal{vol}_m(A_\\varepsilon)}{\\varepsilon^{m-D_M}},\\\\\n\\mathscr{M}_* &:=\\mathscr{M}_*^{D_M}(A)=\\liminf_{\\varepsilon\\to 0^+} \\frac{\\textnormal{vol}_m(A_\\varepsilon)}{\\varepsilon^{m-D_M}}.\n\\end{align*} \nIf the upper and lower Minkowski contents agree and lie in $(0,\\infty)$, then $A$ is said to be \\emph{Minkowski measurable} and the \\emph{Minkowski content of $A$} is given by\n\\begin{align*}\n\\mathscr{M} &:=\\lim_{\\varepsilon\\to 0^+}\\frac{\\textnormal{vol}_m(A_\\varepsilon)}{\\varepsilon^{m-D_M}}.\n\\end{align*} \n\\end{definition}\n\nFor example, if $A$ is such that $\\textnormal{vol}_m(A_\\varepsilon) \\sim M\\varepsilon^{\\alpha}$ as $\\varepsilon \\to 0^+$, then $\\dim_{B}A=m-\\alpha$ and $\\mathscr{M}=M$.\n\n\\begin{openproblem}\nIf $A$ and $B$ are Minkowski measurable in\n$\\mathbf{R}^m$ and $\\mathbf{R}^n$, respectively, is their Cartesian product $A\\times B$ Minkowski measurable in $\\mathbf{R}^{m+n}$\\emph{?} See also Remark \\ref{rmk:Resman} below dealing with the so-called normalized Minkowski content, and its independence of the ambient dimension~$m$. \n\\end{openproblem}\n\n\\begin{remark}\nAnother question to consider is whether or not the union $A\\cup B$ of two Minkowski measurable sets is Minkowski measurable. If not, it would be interesting to find an explicit counter-example. (The answer is clearly affirmative if $A$ and $B$ are a positive distance apart.)\n\\end{remark}\n\n\nThe Minkowski measurable sets on the real line have been characterized in \\cite{LapPo1}; see also Theorem~\\ref{thm:CriterionForMinkowskiMeasurability}. Some classes of Minkowski measurable sets are known in the plane in the case of smooth spirals, see \\cite{ZuZup}, and in the case of discrete spirals, see \\cite{MeSa}. It is interesting that in general, bilipschitz $C^1$ mappings do not preserve Minkowski measurability, even for subsets of the real line; see \\cite{MeSa}.\n\nWe close this section with the following example.\n\n\n\\begin{example}\nIf $A$ is any bounded set in $\\mathbb{R}^m$ such that its closure is of positive $m$-dimensional Lebesgue measure, then $\\dim_BA = m$. Indeed, $\\textnormal{vol}_m(\\overline{A}) > 0$ immediately implies that $\\underline{\\dim}_B\\overline{A} = m$, and therefore by property (iii) of Remark \\ref{rmk:PropertiesBoxCountingDimension}, $\\underline{\\dim}_BA = m$. Since $\\overline{\\dim}_BA \\leq m$, this proves that $\\dim_BA$ exists and $\\dim_BA = m$. In particular, the claim holds for any Lebesgue nonmeasurable set $A\\subset\\mathbb{R}^m$. Indeed, the closure of such a set $A$ cannot be of Lebesgue measure zero (i.e., we cannot have $\\textnormal{vol}_m(\\overline{A}) = 0$) since, in that case, $A$ would also be of Lebesgue measure zero, implying that $A$ is Lebesgue measurable (because Lebesgue measure is complete in this setting).\\footnote{For the notions of measure theory used in this example, we refer, e.g., to \\cite{Cohn}.}\n\\end{example}\n\n\n\\section{Fractal strings and zeta functions}\n\\label{sec:FractalStringsAndZetaFunctions}\n\nIn this section, we discuss a few of the many results on fractal strings presented in \\cite{LapvF6}.\n\n\n\\subsection{Fractal strings and ordinary fractal strings}\n\\label{sec:FractalStringsAndOrdinaryFractalStrings}\n\n\\begin{definition}\n\\label{def:FractalString} \nA \\emph{fractal string} $\\mathcal{L}$ is a nonincreasing sequence of positive real numbers which tends to zero. Hence,\n\\begin{align*}\n\\mathcal{L} &= (\\ell_j)_{j \\in \\mathbb{N}}, \n\\end{align*}\nwhere $(\\ell_j)_{j\\in\\mathbb{N}}$ is nonincreasing and $\\lim_{j \\rightarrow \\infty}\\ell_j=0$. Equivalently, a fractal string $\\mathcal{L}$ can be thought of as the collection \n\\[\n\\{ l_n : l_n \\textnormal{ has multiplicity } m_n, n \\in \\mathbb{N} \\},\n\\] \nwhere $(l_n)_{n \\in \\mathbb{N}}$ is a strictly decreasing sequence of positive real numbers and, for each $n\\in\\mathbb{N}$, $m_n$ is the number of lengths (or, more generally, `scales') $\\ell_j$ such that $\\ell_j=l_n$. \n\nAn \\emph{ordinary fractal string} $\\Omega$ is a bounded open subset of the real line. In that case, we can write $\\Omega$ as the disjoint union of its connected components $I_j$ (i.e., $\\Omega=\\cup_{j=1}^\\infty I_j$), and for each $j\\geq 1$, $\\ell_j$ denotes the length of the interval $I_j$ (i.e., $\\ell_j=|I_j|_1$).\\footnote{Note that without loss of generality, we may assume that $\\ell_1\\geq\\ell_2\\geq\\ldots$, with $\\ell_j\\to 0$ as $j\\to \\infty$. We ignore the trivial case where $\\Omega$ is a finite union of open intervals; see Remark \\ref{rmk:FiniteFractalStrings}.}  \n\\end{definition}\n\n\n\n\\begin{remark}\n\\label{rmk:FiniteFractalStrings}\nIn \\cite{LapvF6}, for instance, finite fractal strings (i.e., nonincreasing sequences of real numbers with a finite number of positive terms) are allowed. However, for reasons described in Remark \\ref{rmk:AbscissaForFiniteSequences}, the finite case is not considered in this paper.\n\\end{remark}\n\nIf, as in Definition \\ref{def:FractalString} above, an ordinary fractal string $\\Omega$ is written as the union of a countably infinite collection of disjoint open intervals $I_j$ (necessarily its connected components), then the lengths $\\ell_j$ of the intervals $I_j$ comprise a fractal string $\\mathcal{L}$. Moreover, $\\dim_M(\\partial\\Omega)$ is given by\n\\begin{align} \n\\label{eqn:OrdinaryFractalStringMinkowski} \\dim_M(\\partial\\Omega)\t&= \\inf\\{\\alpha\\geq 0 : V_{\\textnormal{inner}}(\\varepsilon)= O(\\varepsilon^{1-\\alpha}) \\textnormal{ as } \\varepsilon \\to 0^+\\},\n\\end{align}\nwhere $V_{\\textnormal{inner}}(\\varepsilon)$ is the $1$-dimensional Lebesgue measure of the inner $\\varepsilon$-neighborhood of $\\Omega$ (see formula  \\eqref{eqn:VolumeInnerNeighborhood} in Notation \\ref{not:DistanceVolumeAndBigO}). In fact, Equation \\eqref{eqn:OrdinaryFractalStringMinkowski} is used to define the \\emph{Minkowski dimension} of (the boundary of) an ordinary fractal string in \\cite{LapvF6}.\\footnote{More specifically, $\\dim_M(\\partial\\Omega)$ should really be denoted by $\\dim_{M,\\textnormal{inner}}(\\partial\\Omega)$ and called the \\emph{inner Minkowski dimension} of $\\partial\\Omega$ (or of $\\mathcal{L}$).} Moreover, it is shown in \\cite{LapPo1} that $V_{\\textnormal{inner}}(\\varepsilon)$, and hence also $\\dim_M(\\partial\\Omega)$, depends only on the fractal string $\\mathcal{L}$ (but not on the particular rearrangement of the intervals $I_j$ composing $\\Omega$).\n\n\\begin{definition}\n\\label{def:AbscissaOfConvergence}\nLet $\\mathcal{L}$ be a fractal string. The \\emph{abscissa of convergence} of the Dirichlet series $\\sum_{j=1}^\\infty\\ell_j^s$ is defined by\n\\begin{align}\n\\label{eqn:AbscissaOfConvergence}\n\\sigma\t&=\\inf \\left\\{ \\alpha \\in \\mathbb{R} : \\sum_{j=1}^\\infty\\ell_j^\\alpha<\\infty \\right\\}.\n\\end{align}\nThus, $\\{ s \\in \\mathbb{C} : \\textnormal{Re}(s)>\\sigma \\}$ is the largest open half-plane on which this series converges; see, e.g., \\cite[\\S VI.2]{Ser}.\n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:AbscissaForFiniteSequences}\nIf $\\mathcal{L}$ were allowed to be a finite sequence of positive real numbers (as in \\cite{LapvF6}), then we would have $\\sigma=-\\infty$ since the corresponding Dirichlet series would be an entire function. In the context of this paper, we always have that $\\sigma\\geq 0$ (since $\\sum_{j=1}^\\infty \\ell_j^\\alpha$ is clearly divergent when $\\alpha=0$). This explains why we consider only (bounded) infinite sets in the development of \\emph{box-counting fractal strings} in Section \\ref{sec:BoxCountingFractalStringsAndZetaFunctions}. Indeed, for clarity of exposition, we only consider fractal strings consisting of infinitely many positive lengths (or scales), and hence, ordinary fractal strings comprising infinitely many disjoint intervals. \n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:ConvergenceAndDivergence}\nA key distinction between a fractal string $\\mathcal{L}$ and an ordinary fractal string $\\Omega$ lies in the sum of the corresponding lengths (or scales), denoted $(\\ell_j)_{j\\in\\mathbb{N}}$ in either case. Specifically, since an ordinary fractal string $\\Omega$ is bounded, $\\sum_{j=1}^\\infty \\ell_j$ is necessarily convergent. On the other hand, for a fractal string $\\mathcal{L}$, $\\sum_{j=1}^\\infty \\ell_j$ may be divergent. See Example \\ref{eg:BoxCountingFractalStringOf1DFractal} for a bounded set in $\\mathbb{R}^2$ whose \\emph{box-counting fractal string} is a fractal string whose lengths have an unbounded sum and yet contains pertinent information regarding the bounded set. (In a somewhat different setting, many other classes of examples are provided in \\cite[esp. \\S 13.1 \\& \\S 13.3]{LapvF6} and in \\cite{LapPe2,LapPeWin}.)\n\\end{remark}\n\n\\begin{definition}\n\\label{def:GeometricZetaFunction}\nLet $\\mathcal{L}$ be a fractal string. The \\emph{geometric zeta function} of $\\mathcal{L}$ is defined by \n\\begin{align}\n\\label{eqn:GeometricZetaFunction}\n\\zeta_\\mathcal{L}(s)\t&= \\sum_{j=1}^\\infty\\ell_j^s,\n\\end{align}\nwhere $s \\in \\mathbb{C}$ and  $\\textnormal{Re}(s)>D_\\mathcal{L}:=\\sigma$. The \\emph{dimension} of $\\mathcal{L}$, denoted $D_\\mathcal{L}$, is defined as the abscissa of convergence $\\sigma$ of the Dirichlet series which defines $\\zeta_\\mathcal{L}$.\n\\end{definition}\n\nIn order to define the complex dimensions of a fractal string, as in \\cite{LapvF6}, we assume there exists a meromorphic extension of the geometric zeta function $\\zeta_\\mathcal{L}$ to a suitable region. First, consider the \\emph{screen} $S$ as the contour \n\\begin{align}\n\\label{eqn:Screen}\nS:S(t)+it \\quad &(t \\in \\mathbb{R}),\n\\end{align}\nwhere $S(t)$ is a continuous function $S:\\mathbb{R} \\to [-\\infty,D_\\mathcal{L}]$. Next, consider the \\emph{window} $W$ as the set \n\\begin{align}\n\\label{eqn:Window}\nW\t&= \\{s \\in \\mathbb{C} : \\textnormal{Re}(s) \\geq S(\\textnormal{Im}(s))\\}.\n\\end{align}\n\nBy a mild abuse of notation, we denote by $\\zeta_\\mathcal{L}$ both the geometric zeta function of $\\mathcal{L}$ and its meromorphic extension to some region. \n\n\\begin{definition}\n\\label{def:ComplexDimensions}\nLet $W \\subset \\mathbb{C}$ be a window on an open connected neighborhood of which $\\zeta_\\mathcal{L}$ has a meromorphic extension. The set of (\\emph{visible}) \\emph{complex dimensions} of $\\mathcal{L}$ is the set $\\calD_\\mathcal{L}=\\calD_\\mathcal{L}(W)$ given by\n\\begin{align}\n\\calD_\\mathcal{L}\t&= \\left\\{ \\omega \\in W : \\zeta_\\mathcal{L} \\textnormal{ has a pole at } \\omega \\right\\}.\n\\end{align}\n\\end{definition}\n\n\\noindent In the case where $\\zeta_\\mathcal{L}$ has a meromorphic extension to $W=\\mathbb{C}$, the set $\\calD_\\mathcal{L}$ is referred to as the \\emph{complex dimensions} of $\\mathcal{L}$.  Such is the case for the Cantor string $\\Omega_{CS}$.\n\n\\begin{example}[Complex dimensions of the Cantor string]\n\\label{eg:ComplexDimensionsOfCantorString}\nThe Cantor string $\\Omega_{CS}$ is the ordinary fractal string given by $\\Omega_{CS}=[0,1] \\setminus C$, where $C$ is the Cantor set (see Example \\ref{eg:BoxCountingDimensionOfCantorSet}). The lengths of the Cantor string are given by the fractal string \n\\begin{align}\n\\label{eqn:CantorStringLengths}\n\\mathcal{L}_{CS}\t&=\\{3^{-n}: 3^{-n} \\textnormal{ has multiplicity } 2^{n-1}, n \\in \\mathbb{N}\\}.\n\\end{align}\nThe geometric zeta function of the Cantor string, denoted $\\zeta_{CS}$, is given by\n\\begin{align}\n\\label{eqn:CantorStringGeometricZetaFunction}\n\\zeta_{CS}(s)\t&:= \\zeta_{\\mathcal{L}_{CS}}(s)= \\sum_{n=1}^\\infty 2^{n-1}3^{-ns} = \\frac{3^{-s}}{1-2\\cdot3^{-s}}.\n\\end{align} \nThe closed form on the right-hand side of Equation \\eqref{eqn:CantorStringGeometricZetaFunction} allows for the meromorphic continuation of $\\zeta_{CS}$ to all of $\\mathbb{C}$. Hence, $\\zeta_{CS}=3^{-s}(1-2\\cdot 3^{-s})^{-1}$ for all $s \\in \\mathbb{C}$. It follows that the complex dimensions of the Cantor string, denoted $\\calD_{CS}$, are the complex roots of the Moran equation $2\\cdot3^{-s}=1$. Thus, with the window $W$ chosen to be all of $\\mathbb{C}$ (so that $\\calD_{CS}:=\\calD_{\\mathcal{L}_{CS}}=\\calD_{\\mathcal{L}_{CS}}(\\mathbb{C})$, in the notation of Definition \\ref{def:ComplexDimensions}), we have\n\\begin{align}\n\\label{eqn:CantorStringComplexDimensions}\n\\calD_{CS}\t&=\\left\\{ \\log_3{2} + i\\frac{2\\pi}{\\log 3}z : z \\in \\mathbb{Z} \\right\\}.\n\\end{align}\nNote that by Equation \\eqref{eqn:BoxCountingDimensionOfCantorSet} the dimension $D_{CS}:=D_{\\mathcal{L}_{CS}}=\\log_{3}2$ of the Cantor string coincides with $\\dim_{B}C=\\dim_{M}C$, and this value is the unique real-valued complex dimension of the Cantor string.\n\\end{example}\n\n\n\\subsection{Geometric counting function of a fractal string}\n\\label{sec:GeometricCountingFunctionOfAFractalString}\n\nThe results in this section connect the counting function of the lengths of a fractal string to its dimension and geometric zeta function.\n\n\\begin{definition}\n\\label{def:GeometricCountingFunction}\nThe \\emph{geometric counting function} of $\\mathcal{L}$, or the \\emph{counting function of the reciprocal lengths} of $\\mathcal{L}$, is given by\n\\begin{align*}\n\\displaystyle N_\\mathcal{L}(x) &:= \\#\\{j \\in \\mathbb{N} : \\ell_j^{-1} \\leq x\\} = \\sum_{n \\in \\mathbb{N}, \\, l_n^{-1} \\leq\\, x}m_n\n\\end{align*}\nfor $x>0$.\n\\end{definition}\n\nThe following easy proposition is identical to Proposition 1.1 of \\cite{LapvF6}.\n\n\\begin{proposition}\n\\label{prop:CountingFunctionD}\nLet $\\alpha \\geq 0$ and $\\mathcal{L}$ be a fractal string. Then $N_\\mathcal{L}(x)=O(x^\\alpha)$ as $x \\rightarrow \\infty$ if and only if\n$\\ell_j=O(j^{-1/\\alpha})$ as $j \\rightarrow \\infty$. \n\\end{proposition}\n\n\\begin{proof}\nSuppose that for some $C>0$ we have \n\\[\nN_\\mathcal{L}(x) \\leq Cx^\\alpha.\n\\]\nLet $x=\\ell_j^{-1}$, then $j \\leq C\\ell_j^{-\\alpha}$, which implies that \n\\[\n\\ell_j = O(j^{-1/\\alpha}).\n\\] \nConversely, if $\\ell_j \\leq cj^{-1/\\alpha}$\nfor $j \\in \\mathbb{N}$ and some $c>0$, then given $x>0$, we have \n\\[\n\\ell_j^{-1} > x \\textnormal{ for } j > (cx)^\\alpha.\n\\]\nTherefore, \n\\[\nN_\\mathcal{L}(x) \\leq (cx)^\\alpha.\n\\]\n\\end{proof}\n\n\\begin{remark}\n\\label{rmk:MoreAsymptoticBehavior}\nMany additional (and harder) results connecting the asymptotic behavior of the geometric counting function, the spectral counting function, and the (upper and lower) Minkowski content(s) of a fractal string $\\mathcal{L}$ are provided in \\cite{LapPo1}. The simplest one states that $N_\\mathcal{L}(x)=O(x^\\alpha)$ as $x\\to\\infty$ (i.e., $\\ell_j=O(j^{-1/\\alpha})$ as $j\\to\\infty$) if and only if $\\mathscr{M}^{*\\alpha}(\\partial\\Omega)<\\infty$, where (consistent with our earlier comment) $\\mathscr{M}^{*\\alpha}(\\partial\\Omega)$ is given as in Definition \\ref{def:MinkowskiContent0} except with $\\textnormal{vol}_m(\\cdot)$ replaced with $V_{\\textnormal{inner}}(\\cdot)$.\n\\end{remark}\n\n\\begin{notation}\n\\label{not:D}\nThe infimum of the nonnegative values of $\\alpha$ which satisfy Proposition \\ref{prop:CountingFunctionD} plays a key role in our results. Hence, we let $D_N$ denote that special value. That is,\n\\begin{align}\n\\label{eqn:CountingFunctionD}\nD_N:= \\inf\\{\\alpha \\geq 0 : N_\\mathcal{L}(x)=O(x^\\alpha) \\textnormal{ as } x\\to\\infty\\}.\n\\end{align}\n\\end{notation}\n\nThe following lemma is a restatement of a portion of Lemma 13.110 of \\cite{LapvF6}.\\footnote{In fact, a stronger result holds in the setting of generalized fractal strings (viewed as measures) in \\cite{LapvF6}, but it is beyond the scope of this paper. (See also \\cite{LapLuvF1}.)}\n\n\\begin{lemma}\n\\label{lem:GeometricZetaFunctionIntegralTransformCountingFunction}\nLet $\\mathcal{L}$ be a fractal string. Then \n\\begin{align}\n\\label{eqn:GeometricZetaFunctionIntegralTransformCountingFunction}\n\\zeta_\\mathcal{L}(s)\t&= s\\int_0^\\infty N_\\mathcal{L}(x)x^{-s-1}dx\n\\end{align}\nand, moreover, the integral converges \\emph{(}and hence, Equation \\eqref{eqn:GeometricZetaFunctionIntegralTransformCountingFunction} holds\\emph{)}  if and only if $\\sum_{j=1}^\\infty \\ell_j^s$ converges, i.e., if and only if $\\textnormal{Re}(s)>D_\\mathcal{L}=\\sigma$.\n\\end{lemma}\n\n\\begin{proof}\nLet $s \\geq 0$ be a real number. (The case where $s \\in \\mathbb{C}$ follows immediately from this case by analytic continuation for $\\textnormal{Re}(s)>D_\\mathcal{L}=\\sigma$ since $\\zeta_\\mathcal{L}$ is holomorphic in that half-plane.) For any given $n \\in \\mathbb{N}$, we have  \n\\begin{align*}\ns\\int_0^{\\ell_n^{-1}} N_\\mathcal{L}(x)x^{-s-1}dx\t&= \\sum_{j=1}^{n-1}s\\int_{\\ell_j^{-1}}^{\\ell_{j+1}^{-1}} N_\\mathcal{L}(x)x^{-s-1}dx\n\t= \\sum_{j=0}^{n-1}j(\\ell_j^s-\\ell_{j+1}^s)\n\\end{align*}\nsince $N_\\mathcal{L}(x)=0$ for $x<\\ell_1^{-1}$ and $N_\\mathcal{L}(x)=j$ for $\\ell_j^{-1} \\leq x < \\ell_{j+1}^{-1}$. Furthermore,\n\\begin{align*}\ns\\int_0^{\\ell_j^{-1}} N_\\mathcal{L}(x)x^{-s-1}dx\t&= \n\\sum_{j=1}^{n-1}j\\ell_j^s-\\sum_{j=1}^{n}(j-1)\\ell_j^s\n\t= \\sum_{j=1}^{n}\\ell_j^s-n\\ell_n^s.\n\\end{align*}\nNow, for $s\\geq 0$, we have $n\\ell_n^s \\leq 2\\sum_{j=[n/2]}^n\\ell_j^s$. Thus, Equation \\eqref{eqn:GeometricZetaFunctionIntegralTransformCountingFunction} holds if and only if $\\sum_{j=1}^\\infty \\ell_j^s$ converges.\\footnote{That is, if and only if $s>D_\\mathcal{L}=\\sigma$ (or, more generally, if $s\\in\\mathbb{C}$, if and only if $\\textnormal{Re}(s)>D_\\mathcal{L}$).} Moreover,\n\\begin{align*}\n\\lim_{n \\rightarrow \\infty} s\\int_0^{\\ell_n^{-1}} N_\\mathcal{L}(x)x^{-s-1}dx\t&= \\zeta_\\mathcal{L}(s) \n\\end{align*}\nsince the tail $\\sum_{j=[n/2]}^\\infty \\ell_j^s$ converges to zero. (Here, $[y]$ denotes the integer part of the real number $y$.)\n\\end{proof}\n\nThe following proposition will be used to prove a portion of our main result, Theorem \\ref{thm:MainResult} (cf. Theorem 13.111 and Corollary 13.112 of \\cite{LapvF6}, as well as  \\cite{LapLuvF1,LapLuvF2}, where this proposition is established in the context of $p$-adic fractal strings and also of ordinary (real) fractal strings).\n\n\\begin{proposition}\n\\label{prop:CountingFunctionDFractalStringD}\nLet $\\mathcal{L}$ be a fractal string. Then \n\\begin{align*}\nD_\\mathcal{L}\t&=D_N,\n\\end{align*}\nwhere $D_\\mathcal{L}=\\sigma$ is the dimension of $\\mathcal{L}$ given by Equation \\eqref{eqn:AbscissaOfConvergence} \\emph{(}and Definition \\ref{def:GeometricZetaFunction}\\emph{)} and $D_N$ is given by Equation \\eqref{eqn:CountingFunctionD}.\n\\end{proposition}\n\n\\begin{proof}\nThe proof given here follows that of \\cite{LapvF6}, \\emph{loc.~cit.} (See also \\cite{LapLuvF1}.) First, suppose $\\textnormal{Re}(s)>D_N$. Denoting $t=\\textnormal{Re}(s)$, we choose any fixed  $\\alpha\\in(D_N,t)$.\n Using Lemma \\ref{lem:GeometricZetaFunctionIntegralTransformCountingFunction}, for $x_1=(\\ell_1)^{-1}$ we have\n\\begin{align*}\n|\\zeta_\\mathcal{L}(s)|\n\t&\\leq |s|\\int_{x_1}^\\infty Cx^{\\alpha}x^{-t-1}dx = \\left[\\frac{|s|Cx^{\\alpha-t}}{\\alpha-t}\\right]_{x_1}^\\infty \n\t= 0 - \\frac{|s|Cx_1^{\\alpha-t}}{\\alpha-t},\n\\end{align*}\nsince $\\alpha-t<0$. Hence, $|\\zeta_\\mathcal{L}(s)|<\\infty$. In other words,  $t>D_{\\mathcal{L}}$ for any $t>D_N$. Letting $t\\searrow D_N$, we obtain that $D_{\\mathcal{L}}\\le D_N$.\n\n\nFor the converse, suppose $\\alpha < D_N$. Then $N_\\mathcal{L}(x)$ is not $O(x^\\alpha)$ as $x \\rightarrow \\infty$. So, there exists a strictly increasing sequence $(x_j)_{j\\in\\mathbb{N}}$ with $x_1 \\geq \\ell_1^{-1}$ which tends to infinity such that \n\\[\nN_\\mathcal{L}(x) \\geq jx_j^\\alpha\n\\]\nfor each $j$. Then, for $t \\leq 1$, \n\\[\nt\\int_{\\ell_1^{-1}}^\\infty N_\\mathcal{L}(x)x^{-t-1}dx \\geq \\sum_{j=1}t\\int_{x_j}^{x_{j+1}}jx^\\alpha x^{-t-1}dx\n\\]\nsince $N_\\mathcal{L}(x)$ is increasing.\n\nWe estimate $t\\int_{x_j}^{x_{j+1}}x^{-t-1}dx \\geq x_j^{-t}$ to obtain\n\\[\nt\\int_{\\ell_1^{-1}}^\\infty N_\\mathcal{L}(x)x^{-t-1}dx \\geq \\sum_{j=1}^\\infty jx_j^{\\alpha-t}.\n\\]\nFor $t \\leq \\alpha$, the sum diverges. Hence, $D_\\mathcal{L} \\geq \\alpha$ for all $\\alpha<D_N$, and so $D_\\mathcal{L} \\geq D_N$.\n\\end{proof}\n\n\n\nFor a given fractal string $\\mathcal{L}$, Theorem \\ref{thm:CountingFunctionAndComplexDimensions} (cf. Theorem 5.10 and Theorem 5.18 in \\cite{LapvF6}) shows that under mild conditions the complex dimensions $\\calD_\\mathcal{L}$ contain enough information to determine the geometric counting function $N_\\mathcal{L}$ (at least, asymptotically, and in important geometric situations, exactly\\footnote{This is the case, for instance, for self-similar strings.}). \n\n\n\\begin{theorem}\n\\label{thm:CountingFunctionAndComplexDimensions}\nLet $\\mathcal{L}$ be a fractal string such that $\\calD_\\mathcal{L}$ consists entirely of simple poles with respect to a window $W$. Then, under certain mild growth conditions on $\\zeta_\\mathcal{L}$,\\footnote{Namely, if $\\zeta_\\mathcal{L}$ is \\emph{languid} (see \\cite[Def.~5.2]{LapvF6}) of a suitable order.} we have\n\\begin{align}\n\\label{eqn:CountingFunctionAndComplexDimensions}\nN_\\mathcal{L}(x) = \\sum_{\\omega \\in \\calD_\\mathcal{L}}\n\\frac{x^{\\omega}}{\\omega}\\textnormal{res}(\\zeta_\\mathcal{L}(s);\\omega) + \\{\\zeta_\\mathcal{L}(0)\\} + R(x), \n\\end{align}\nwhere $R(x)$ is an error term of small order and the term in braces is included only if $0 \\in W\\backslash \\calD_\\mathcal{L}$.\n\\end{theorem}\n\n\\begin{remark}\n\\label{rmk:SimpleAndStronglyLanguid}\nIf the poles are not simple, the explicit formula for $N_{\\mathcal{L}}$ is slightly more complicated (see\\cite[Chs.~5,6]{LapvF6}). If an ordinary fractal string $\\Omega$ is {\\it strongly languid} (see \\cite[Def.~5.3]{LapvF6}), then by Theorem 5.14 and Theorem 5.22 of \\cite{LapvF6}, Equation \\eqref{eqn:CountingFunctionAndComplexDimensions} holds with no error term (i.e., $W=\\mathbb{C}$ and $R(x) \\equiv 0$) and hence, formula \\eqref{eqn:CountingFunctionAndComplexDimensions} is exact in that case.\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:FractalTubeFormulas}\nSimilar (but harder to derive) explicit formulas called \\emph{fractal tube formulas} are obtained in \\cite[Ch.~8]{LapvF6}, which, as described therein, allows for the expression of $V_\\textnormal{inner}(\\varepsilon)$ in terms of the underlying (visible) complex dimensions of $\\mathcal{L}$. (Still in \\cite{LapvF6}, they are used, in particular, to derive the equivalence of (i) and (iii) in Theorem \\ref{thm:CriterionForMinkowskiMeasurability} below.) We will return to this topic in Section \\ref{sec:SummaryOfResultsAndOpenProblems} when discussing Open Problems \\ref{op:BoxCountingAndCountingFormulas} and \\ref{op:VolumeTubularNeighborhood}.\n\\end{remark}\n\nAnalogous results regarding connections between the structure of the complex dimensions $\\calD_\\mathcal{L}$ of an ordinary fractal string $\\Omega$ with lengths $\\mathcal{L}$ and the (inner) Minkowski measurability of $\\partial\\Omega$ are presented in the next section.\n\n\n\\subsection{Classic results}\n\\label{sec:ClassicResults}\n\nThe following theorem is precisely Theorem 1.10 of \\cite{LapvF6}. It is actually a consequence of a classic theorem of Besicovitch and Taylor (see \\cite{BesTa}) stated in terms of ordinary fractal strings, and was first observed in this context in \\cite{Lap2}.\\footnote{There is, however, one significant difference with the setting of \\cite{BesTa}. Namely, here, as in \\cite{Lap2} and \\cite{LapvF6}, we are assuming that we are working with the \\emph{inner} (rather than ordinary) Minkowski dimension and Minkowski content of $\\partial\\Omega$; see the statement and the proof of Theorem 1.10 in \\cite{LapvF6}, along with Equation \\eqref{eqn:VolumeInnerNeighborhood} above. By contrast, in the context of \\cite{BesTa}, one should assume that $\\Omega$ is of full measure in its closed convex hull (i.e., in the smallest compact interval containing it).} \n\n\\begin{theorem}\n\\label{thm:BesicovitchAndTaylor}\nSuppose $\\Omega$ is an ordinary fractal string with infinitely many lengths denoted by $\\mathcal{L}$. Then the abscissa of convergence of $\\zeta_\\mathcal{L}$ coincides with the Minkowski dimension of $\\partial\\Omega$. That is, $D_\\mathcal{L}=\\dim_M(\\partial\\Omega)$.\n\\end{theorem}\n\nThe following result is Theorem 8.15 of \\cite{LapvF6}. For complete details regarding connections between complex dimensions and Minkowski measurability, see \\cite[Ch.~8]{LapvF6}.\n\n\\begin{theorem}[Criterion for Minkowski measurability]\n\\label{thm:CriterionForMinkowskiMeasurability}\nLet $\\Omega$ be an ordinary fractal string whose geometric zeta function $\\zeta_\\mathcal{L}$ has a meromorphic extension which satisfies certain mild growth conditions.\\footnote{Specifically, $\\zeta_\\mathcal{L}$ is languid for a screen $S$ passing strictly between the vertical line $\\textnormal{Re}(s)=D_\\mathcal{L}$ and all the complex dimensions (of the corresponding fractal string) $\\mathcal{L}$ with real part strictly less than $D_\\mathcal{L}$, and not passing through 0.} Then the following are equivalent\\emph{:}\n\\begin{enumerate}\n\t\\item $D_\\mathcal{L}$ is the only complex dimension with real part $D_\\mathcal{L}$, and it is simple.\n\t\\item $N_\\mathcal{L}(x)=cx^{D_\\mathcal{L}}+o(x^{D_\\mathcal{L}})$ as $x \\to\\infty$, for some positive constant $c$.\\footnote{In the spirit of Proposition \\ref{prop:CountingFunctionD}, condition (ii) is easily seen to be equivalent to \n\\[\n\\ell_j=Lj^{-1/D_\\mathcal{L}}+o(j^{-1/D_\\mathcal{L}}) \\quad \\textnormal{as} \\quad j\\to\\infty,\n\\]\nfor some positive constant $L$. In that case, we have $c=L^{D_\\mathcal{L}}$.}\n\t\\item $\\partial\\Omega$ is Minkowski measurable.\n\\end{enumerate}\nMoreover, if any of these conditions is satisfied, then the Minkowski content $\\mathscr{M}$\nof $\\partial\\Omega$ is given by\n\\begin{align*}\n\\mathscr{M}\t&= \\frac{c2^{1-D_\\mathcal{L}}}{1-D_\\mathcal{L}} = 2^{1-D_\\mathcal{L}}\\frac{\\textnormal{res}(\\zeta_\\mathcal{L}(s);D_\\mathcal{L})}{D_\\mathcal{L}(1-D_\\mathcal{L})}. \n\\end{align*}\n\\end{theorem}\n\n\\begin{remark}\n\\label{rmk:TheoryOfComplexDimensions}\nWe note that the equivalence of (ii) and (iii) in Theorem \\ref{thm:CriterionForMinkowskiMeasurability} was first established in \\cite{LapPo1} for any ordinary fractal string, without any hypothesis on the growth of the associated geometric zeta function. As was alluded to in Remark \\ref{rmk:FractalTubeFormulas}, however, the equivalence of (i) and (iii) in Theorem \\ref{thm:CriterionForMinkowskiMeasurability} was proved in \\cite{LapvF6} (and in earlier works of the authors of \\cite{LapvF6}) by using a suitable generalization of Riemann's explicit formula that is central to the theory of complex dimensions and is obtained in \\cite[Chs.~5 \\& 8]{LapvF6}.\n\\end{remark}\n\n\\begin{example}[The Cantor set is not Minkowski measurable]\n\\label{eg:CantorSetNotMinkowskiMeasurable}\nBy Equation \\eqref{eqn:CantorStringComplexDimensions} in Example \\ref{eg:ComplexDimensionsOfCantorString}, there is an infinite collection of complex dimensions $\\omega \\in \\calD_{CS}$ of the Cantor string with real part $D_{CS}=\\log_{3}2$. Hence, by Theorem \\ref{thm:CriterionForMinkowskiMeasurability}, the Cantor set $C$ is not Minkowski measurable. This fact was first established in \\cite{LapPo1} by using the equivalence of (ii) and (iii) and showing that (ii) does not hold. Actually, still in \\cite{LapPo1}, for $\\alpha=\\dim_{B}C=D_{CS}$, both $\\mathscr{M}^{\\alpha*}=\\mathscr{M}^*$ and $\\mathscr{M}^{\\alpha}_*=\\mathscr{M}_*$ are explicitly computed and shown to be different (with $0<\\mathscr{M}_*<\\mathscr{M}^*<\\infty$). This result was significantly refined and extended in \\cite[Ch.~10]{LapvF6} in the broader context of generalized Cantor strings.\n\\end{example}\n\n\\begin{remark}\n\\label{rmk:LatticeNonLattice}\nExample \\ref{eg:CantorSetNotMinkowskiMeasurable} is indicative of another result from \\cite{LapvF6} pertaining to a dichotomy in the properties of self-similar attractors of certain iterated function systems on compact intervals. Specifically, if an iterated function system on a compact interval $I$ satisfies the open set condition with at least one gap and there is some $0<r<1$ and positive integers $k_j$ such that $\\gcd(k_1,\\ldots,k_N)=1$ and the scaling ratios satisfy $r_j=r^{k_j}$ for each $j = 1,\\ldots, N$, then the complement $I \\setminus A$ of the resulting attractor $A$ is an ordinary fractal string known as a \\emph{lattice self-similar string}. For example, the Cantor string $\\Omega_{CS}=[0,1] \\setminus C$ is a lattice self-similar string. If no such $r$ exists, then $I \\setminus A$ is a \\emph{nonlattice self-similar string}. The complex dimensions of a self-similar string are given by (a subset of) the complex roots of the corresponding Moran equation \\eqref{eqn:Moran}. In the lattice case there are countably many complex dimensions with real part $D_\\mathcal{L}=\\dim_{B}A=\\dim_{M}A$, so by Theorem \\ref{thm:CriterionForMinkowskiMeasurability}, $A$ is not Minkowski measurable. In the nonlattice case, Theorem \\ref{thm:CriterionForMinkowskiMeasurability} does not necessarily apply (because its hypotheses need not be satisfied, see \\cite[Example~5.32]{LapvF6}), however the only complex dimension with real part $D_\\mathcal{L}$ is $\\dim_{B}A=\\dim_{M}A$ and by Theorem 8.36 of \\cite{LapvF6} we have that $A$ is Minkowski measurable. Therefore, the boundary of a self-similar string is Minkowski measurable if and only if it is nonlattice. See \\cite[\\S 8.4]{LapvF6} for details.\n\\end{remark}\n \nWe conclude the section on classic results with the following remark which, in light of the expression for $V_{\\textnormal{inner}}(\\varepsilon)$ obtained in \\cite{LapPo1} (see also \\cite[Eq.~(8.1)]{LapvF6}), can be deduced from Lemma 1 of \\cite[\\S 1.4]{Lev}.\\footnote{For convenience, Remark \\ref{rmk:DimensionEqualsOrder}\nis stated in the language of fractal strings. A direct (and independent) proof of the fact that Equation \\eqref{eqn:DimensionEqualsOrder} holds \ncan be found in \\cite{LapvF6}. (See also \\cite{LapLuvF1}.) Also, Equation \\eqref{eqn:DimensionEqualsOrder} is a simple consequence of results obtained in \\cite{LapPo1}.} The remark below provides yet another connection between counting functions and dimensions, although it is a simple restatement of Proposition \\ref{prop:CountingFunctionDFractalStringD} in our context.\n\n\\begin{remark}\n\\label{rmk:DimensionEqualsOrder}\nFor a fractal string $\\mathcal{L}$, the \\emph{order of the geometric counting function} $N_\\mathcal{L}$, denoted $\\rho_\\mathcal{L}$, is given by \n\\begin{align}\n\\label{eqn:OrderOfCountingFunction}\n\\rho_\\mathcal{L}\t&:=\\limsup_{x \\rightarrow \\infty} \\frac{\\log{N_\\mathcal{L}(x)}}{\\log{x}}.\n\\end{align}  \nWe have that the dimension $D_\\mathcal{L}$ coincides with $\\rho_\\mathcal{L}$. That is,\n\\begin{align}\n\\label{eqn:DimensionEqualsOrder}\nD_\\mathcal{L}\t&=\\rho_\\mathcal{L}.\n\\end{align}\nNote that, for a given fractal string $\\mathcal{L}$, the order of the counting function $\\rho_\\mathcal{L}$ given in Equation \\eqref{eqn:OrderOfCountingFunction} and the value $D_N$ given in Equation \\eqref{eqn:CountingFunctionD} provide essentially the same information regarding the geometric counting function $N_\\mathcal{L}$. Indeed, it can be shown directly that $\\rho_\\mathcal{L}=D_N$, and hence Equation \\eqref{eqn:DimensionEqualsOrder} would follow from Proposition \\ref{prop:CountingFunctionDFractalStringD}. This connection is examined further in \\cite{LapRoZu}.\n\\end{remark}\n\nIn the next section, motivated by the box-counting function $N_B$ and connections between the geometric counting function $N_\\mathcal{L}$ and dimension $D_\\mathcal{L}$ of a fractal string $\\mathcal{L}$, we define and investigate the properties of box-counting fractal strings.\n\n\n\\section{Box-counting fractal strings and zeta functions}\n\\label{sec:BoxCountingFractalStringsAndZetaFunctions}\n\nIn this section, we develop the definition of and results pertaining to \\emph{box-counting fractal strings}. These fractal strings are defined in order to provide a framework in which one may, perhaps, extend the results on ordinary fractal strings via associated zeta functions and complex dimensions in \\cite{LapvF6} to bounded sets. Further exploration with box-counting fractal strings, such as Minkowski measurability of bounded sets, is central to the development of the authors' paper \\cite{LapRoZu}. The box-counting fractal string and the box-counting zeta function for bounded sets in Euclidean spaces were introduced by the second author during the First International Meeting of the Permanent International Session of Research Seminars (PISRS) at the University of Messina, PISRS Conference 2011: Analysis, Fractal Geometry, Dynamical Systems, and Economics.\\footnote{See Remark \\ref{rmk:PrecursorToBoxCountingZetaFunction}.} The introduction took place after listening to a lecture of the third author about his results (with the first author and Goran Radunovi\\'c) in \\cite{LapRaZu} on distance and tube zeta functions for arbitrary compact subsets of $\\mathbb{R}^m$. Some of these results are also discussed in Section \\ref{sec:DistanceAndTubeZetaFunctions} below.\n\n\\begin{remark}\n\\label{rmk:PrecursorToBoxCountingZetaFunction}\nAt the time, the second author did not yet know that the first author had already proposed (since the early 2000s, in several research documents) to introduce and study a `box-counting zeta function' (defined via the Mellin transform of a `box-counting function'), much as in Corollary \\ref{cor:EquivalentZetaFunctions} and Definition \\ref{def:BoxCountingFunction} as well as Remark \\ref{rmk:VariousCountingFunctions}) in order to develop a higher-dimensional theory of complex dimensions and obtain the associated fractal tube formulas for suitable fractal subsets of $\\mathbb{R}^m$ (established when $m=1$ in \\cite[Ch.~8]{LapvF6}). The explicit construction of a `box-counting fractal string' associated with a given bounded set $A\\subset\\mathbb{R}^m$ is new and potentially quite useful, however.\\footnote{Compare, when $m=1$, the construction provided in \\cite{LapMa} of certain fractal strings associated with suitable monotonically increasing step functions, viewed as `geometric counting functions' (in the sense of \\cite{LapPo1,LapvF6} and of Definition \\ref{def:GeometricCountingFunction} above).} Of course, the other results of \\cite{LapRoZu} described in Section \\ref{sec:BoxCountingFractalStringsAndZetaFunctions} are new as well.\n\\end{remark}\n\n\n\\subsection{Definition of box-counting fractal strings}\n\\label{sec:DefinitionOfBoxCountingFractalStrings}\n\nIf $A \\subset \\mathbb{R}^m$ is bounded, then the diameter of $A$, denoted $\\diam(A)$, is finite. So for nonempty $A$ and all $x$ small enough, we have $N_B(A,x)=1$ when $N_B(A,\\cdot)$ is given as in Definition \\ref{def:BoxCountingFunction} or one of the options in Remark \\ref{rmk:VariousCountingFunctions}. Indeed, for a given bounded infinite set $A$, each such box-counting function uniquely defines a fractal string $\\mathcal{L}_B$, which is introduced below and called the \\emph{box-counting fractal string}, by uniquely determining a sequence of distinct scales $(l_n)_{n\\in\\mathbb{N}}$ along with corresponding multiplicities $(m_n)_{n\\in\\mathbb{N}}$. \n\nGiven a fixed bounded infinite set $A$, the range of a chosen box-counting function $N_B(A,\\cdot)$ can be thought of as a strictly increasing sequence of positive integers $(M_n)_{n\\in\\mathbb{N}}$. In this context, we can readily define a fractal string $\\mathcal{L}_B$ whose geometric counting function $N_{\\mathcal{L}_B}$ essentially coincides with $N_B(A,\\cdot)$; see Lemma \\ref{lem:CountingFunctionCorrelation} below. To this end, the key idea is to make the distinct scales $l_n$ of the desired (box-counting) fractal string $\\mathcal{L}_B$ correspond to the scales at which the box-counting function $N_B(A,\\cdot)$ jumps. Furthermore, the multiplicities $m_n$ are defined in order to have the resulting geometric counting function $N_{\\mathcal{L}_B}$ (nearly) coincide with the chosen box-counting function $N_B(A,\\cdot)$. Such \\emph{box-counting fractal strings} potentially allow for the development of a theory of complex dimensions of fractal strings, as presented in \\cite{LapvF6}, by means of results in Section \\ref{sec:FractalStringsAndZetaFunctions} similar to Theorem \\ref{thm:CountingFunctionAndComplexDimensions} above.\\footnote{Parts of such a theory are now developed in \\cite{LapRaZu}, using distance and tube zeta functions; see Section \\ref{sec:DistanceAndTubeZetaFunctions} below for a few sample results. Once the two parallel theories are more fully developed, a challenging problem will consist in comparing and contrasting their respective results and scopes.} These concepts are central to the development of the paper \\cite{LapRoZu}.\n\n\\begin{definition} \n\\label{def:BoxCountingFractalString} \nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$ and let $N_B(A,\\cdot)$ denote a box-counting function given by one of the options in Remark \\ref{rmk:VariousCountingFunctions}. Denote the range of $N_B(A,\\cdot)$ as a strictly increasing sequence of positive integers $(M_n)_{n\\in\\mathbb{N}}$. \nFor each $n\\in\\mathbb{N}$, let $l_n$ be the scale given by\n\\begin{align}\n\\label{eqn:LengthsOfABoxCountingFractalString}\nl_n:=(\\sup\\{x \\in (0,\\infty): N_B(A,x)=M_n\\})^{-1}.\n\\end{align}\nAlso, let $m_1:=M_2$, and for $n\\geq 2$, let $m_n:=M_{n+1}-M_n$. \nThe \\emph{box-counting fractal string} of A, denoted $\\mathcal{L}_B$, is given by\n\\begin{align*}\n\\mathcal{L}_B &:= \\{l_n : l_n \\textnormal{ has multiplicity } m_n, n \\in \\mathbb{N}\\}.\n\\end{align*} \n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:CountingFunctionDefinesUniqueFractalString}\nNote that the distinct scales $l_n$ and the multiplicities $m_n$ are \\emph{uniquely defined by the box-counting function} $N_B(A,\\cdot)$ since $N_B(A,x)$ is nondecreasing as $x\\rightarrow\\infty$. Also, each $l_n$ is equivalently given by\n\\[ \nl_n = \\inf\\{\\varepsilon \\in (0,\\infty): N_B(A,\\varepsilon^{-1})=M_n\\}.\n\\]\n\\end{remark}\n\nIt remains to show that $\\mathcal{L}_B$ is indeed a fractal string; see Definition \\ref{def:FractalString}. That is, since we want to use as many of the results from \\cite{LapvF6} as possible (some of which are presented in Section \\ref{sec:FractalStringsAndZetaFunctions}), we must verify that $\\mathcal{L}_B=(\\ell_j)_{j\\in\\mathbb{N}}$ is a nonincreasing sequence of positive real numbers which tends to zero. This is accomplished with the following proposition, in which other behaviors of $N_B(A,\\cdot)$ are also determined. \n\nFor clarity of exposition and in order to ease the notation used in this section, in particular in the following proposition, take $N_B(A,\\cdot)$ to be defined by option (i) of Remark \\ref{rmk:VariousCountingFunctions} and let $N_B(A,0):=0$. (Completely analogous results hold when $N_B(A,\\cdot)$ is given by Definition \\ref{def:BoxCountingFunction} or one of the other options in Remark \\ref{rmk:VariousCountingFunctions}, \\emph{mutatis mutandis}.) Note that we have $N_B(A,x)\\leq N_B(A,y)$ whenever $0<x<y$. In this setting, and following Remark \\ref{rmk:varepsilonVSx}, $x^{-1}$ denotes the diameter of the sets used to cover $A$. Furthermore, let $x_n:=l_n^{-1}$ for each $n\\in\\mathbb{N}$, and note that we have $N_B(A,x_2)=m_1=M_2$ and \n\\[\nN_B(A,x_{n+1})-N_B(A,x_n)=m_n=M_{n+1}-M_n, \\quad \\textnormal{for } n \\geq 2.\n\\]\n\n\\begin{proposition}\n\\label{prop:BoxCountingJumps}\nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$ and let $l_n$ be given by Equation \\emph{\\eqref{eqn:LengthsOfABoxCountingFractalString}}. Then the sequence $(x_n)_{n\\in\\mathbb{N}}:=(l_n^{-1})_{n\\in\\mathbb{N}}$ is a countably infinite, strictly increasing sequence of positive real numbers such that, for each $n \\in \\mathbb{N}$ and all $x$ such that $x_{n-1} < x \\leq x_n$ \\emph{(}letting $x_0=0$\\emph{)}, we have\n\\begin{align}\n\\label{eqn:BoxCountingJumps}\nN_B(A,x_{n-1}) &< N_B(A,x) = N_B(A,x_n).\n\\end{align}\nFurthermore, \n\\begin{enumerate}\n\\setlength{\\itemsep}{0in}\n\\item $x_1>0$ and $N_B(A,x_1)=1$,\n\\item $x_n \\nearrow \\infty$ as $n \\rightarrow \\infty$, and\n\\item $\\displaystyle \\bigcup_{n \\in \\mathbb{N}} \\{N_B(A,x_n)\\} = \\textnormal{range}~N_B(A,\\cdot)$.  \n\\end{enumerate}\n\\end{proposition}\n\n\\begin{proof}\nWe have that $N_B(A,x)$ is nondecreasing for $x>0$. Further, the range of $N_B(A,\\cdot)$, denoted $\\textnormal{range}~N_B(A,\\cdot)$ (and also realized as the sequence $(M_n)_{n\\in\\mathbb{N}}$ above), is at most countable since it is a subset of $\\mathbb{N}$. In fact, $\\textnormal{range}~N_B(A,\\cdot)$ is countably infinite (otherwise, $A$ would be finite). Hence, $(x_n)_{n \\in \\mathbb{N}}$ is a unique, countably infinite, strictly increasing sequence of positive real numbers such that, for each $n \\in \\mathbb{N}$ and all $x$ such that  $x_{n-1} < x \\leq x_n$ (letting $x_0=0$), we have\n\\begin{align*}\nN_B(A,x_{n-1}) &< N_B(A,x) = N_B(A,x_n).\n\\end{align*}\n\nSince $A$ is bounded and contains more than two elements, there exists a unique $x'\\in (0,\\infty)$ such that $N_B(A,x)=1$ if $0<x \\leq x'$, and $N_B(A,x)>1$ if $x>x'$. By the definition of the sequence $(x_n)_{n\\in\\mathbb{N}}$, we have $x'=x_1$.\n\nNow, suppose $(x_n)_{n\\in\\mathbb{N}}$ has an accumulation point at some $x'' \\in (0,\\infty)$. Then $N_B(A,x'')=\\infty$ since $N_B(A,\\cdot)$ increases by some positive integer value at $x_n$ for each $n \\in \\mathbb{N}$ and since $\\textnormal{range}~N_B(A,\\cdot)\\subset\\mathbb{N}$. However, this contradicts the boundedness of $A$. Further, assuming $N_B(A,\\cdot)$ is bounded implies that $A$ is finite. Hence, $x_n \\nearrow \\infty$ as $n \\rightarrow \\infty$.\n\nLastly, suppose there exists $k \\in \\textnormal{range}~N_B(A,\\cdot)$ such that we have $k \\neq N_B(A,x_n)$ for all $n \\in \\mathbb{N}$. Since $x_n \\nearrow \\infty$ as $n \\rightarrow \\infty$ and $N_B(A,\\cdot)$ is nondecreasing, there exists a unique $n_0 \\in \\mathbb{N}$ such that $x_{n_0-1}<y<x_{n_0}$ for all $y$ such that $N_B(A,y)=k$. However, Equation \\eqref{eqn:BoxCountingJumps} implies $N_B(A,y)=k=N_B(A,x_{n_0})$, which is a contradiction. Therefore, $\\bigcup_{n \\in \\mathbb{N}}\\{N_B(A,x_n)\\} = \\textnormal{range}~N_B(A,\\cdot)$.\n\\end{proof}\n\n\\begin{remark}\n\\label{rmk:BoxCountingFractalStringIsAFractalString}\nBy line (ii) of Proposition \\ref{prop:BoxCountingJumps}, $l_n \\searrow 0$ as $n \\rightarrow \\infty$ and, hence, $\\mathcal{L}_B$ is indeed a fractal string in the sense of Definition \\ref{def:FractalString}.\n\\end{remark}\n\n\n\\begin{example}[Box-counting fractal string of the Cantor set]\n\\label{eg:CantorSetBoxCountingFractalString}\nConsider the Cantor set $C$. For $x>0$, let the box-counting function $N_B(C,x)$ be the minimum number of sets of diameter $x^{-1}$ required to cover $C$ (i.e., as in option (i) of Remark \\ref{rmk:VariousCountingFunctions}). Then the box-counting fractal string $\\mathcal{L}_B$ of $C$ is given by\n\\begin{align}\n\\label{eqn:CantorSetBoxCountingFractalString}\n\\mathcal{L}_B\t&= \\{l_1=1: m_1=2\\}\\cup\\{l_n=3^{-(n-1)}: m_n =2^{n-1}, n \\geq 2 \\}.\n\\end{align}\nIndeed, for each $n \\in \\mathbb{N}$, exactly $2^{n}$ intervals of diameter $3^{-n}$ are required to cover $C$. If $x^{-1}<3^{n}$, then more than $2^{n}$ intervals of diameter $x^{-1}$ are required to cover $C$. \n\\end{example}\n\n\\begin{example}[Box-counting fractal string of a 1-dimensional fractal]\n\\label{eg:BoxCountingFractalStringOf1DFractal}\nConsider the self-similar set $F$ which is the attractor of the IFS $\\mathbf{\\Phi}_1=\\{\\Phi_j\\}_{j=1}^4$ on the unit square $[0,1]^2\\subset\\mathbb{R}^2$ given by\n\\begin{align*}\n\\Phi_1(x)\t&=\\frac{1}{4}x, \\quad \\Phi_2(x)\t=\\frac{1}{4}x+\\left(\\frac{3}{4},0\\right), \\quad\n\\Phi_3(x)\t=\\frac{1}{4}x+\\left(\\frac{3}{4},\\frac{3}{4}\\right), \\quad \\textnormal{and}\\\\ \\Phi_4(x)\t&=\\frac{1}{4}x+\\left(0,\\frac{3}{4}\\right).\n\\end{align*}\nThe Moran equation of $F$ is simply $4\\cdot 4^{-s}=1$, hence $D_{\\mathbf{\\Phi}_1}=\\dim_BF=\\dim_MF=1$ and $F$ is a 1-dimensional self-similar set which is totally disconnected.\n\nLet $N_B(F,x)$ be as in Definition \\ref{def:BoxCountingFunction}. Then, for $x \\in (0,2]$, we have\n\\[\nN_B(F,x)=\n\\begin{cases}\n1, & 0 < x \\leq 2/\\sqrt{2}, \\\\\n2, & 2/\\sqrt{2} < x \\leq 8/\\sqrt{17}, \\\\\n3, & 8/\\sqrt{17} < x \\leq 2.\n\\end{cases}\n\\]\nIndeed, we have: $\\sqrt{2}$ is the distance between $(0,0)$ and $(1,1)$; $\\sqrt{17}/4 $ is the minimum distance between $(0,0),(1,1/4),$ and $(1/4,1)$; and the minimum distance between $(0,0),(0,1),(1,0),$ and $(1,1)$ is $1$. Hence, $M_1=1, M_2=2$, and $M_3=3$. \n\nFor $x>2$, the self-similarity of $F$ implies that $M_n=j4^k$, where $n$ is uniquely expressed as $n=3k+j$ with $k\\in \\mathbb{N}\\cup\\{0\\}$ and $j\\in\\{1,2,3\\}$. So, for $n\\geq 2$, we have $m_n=M_{n+1}-M_n=4^k$ and therefore the box-counting fractal string $\\mathcal{L}_B=(\\ell_j)_{j=1}^\\infty$ of $F$ is the sequence obtained by putting the following collection of distinct scales in nonincreasing order and listing them according to multiplicity:\n\\begin{align*}\n&\\left\\{\\sqrt{2}/2: \\textnormal{ multiplicity } 2 \\right\\}\\\\ \n&\\cup \\left\\{\\sqrt{2}/(2\\cdot 4^k): \\textnormal{ multiplicity } 4^k, k\\in \\mathbb{N} \\right\\}\\\\\n&\\cup \\left\\{\\sqrt{17}/(8\\cdot 4^k): \\textnormal{ multiplicity } 4^k, k\\in \\mathbb{N}\\cup\\{0\\} \\right\\}\\\\\n&\\cup \\left\\{1/(2\\cdot 4^k): \\textnormal{ multiplicity } 4^k, k\\in \\mathbb{N}\\cup\\{0\\} \\right\\}. \n\\end{align*}\n\\end{example}\n\nExamples \\ref{eg:CantorSetBoxCountingFractalString} and \\ref{eg:BoxCountingFractalStringOf1DFractal} will be revisited and expanded upon in the following subsection.\n\n\n\\subsection{Box-counting zeta functions}\n\\label{sec:BoxCountingZetaFunctions}\n\nSuppose $A$ is a bounded infinite subset of $\\mathbb{R}^m$. Each scale $l_n \\in \\mathcal{L}_B$ is distinct and, for $n \\geq 2$, counted according to the multiplicity $m_n:=N_B(A,x_{n+1})-N_B(A,x_n)$. It will help to note that we can also consider $\\mathcal{L}_B$ to be given by the nonincreasing sequence $(\\ell_j)_{j \\in \\mathbb{N}}$, where the distinct values among the $\\ell_j$'s repeat the $l_n$'s according to the multiplicities $m_n$. (The convention of distinguishing the notation $\\ell_j$ and $l_n$ in this way is established in \\cite{LapvF6} and its predecessors, where the distinction allows for various results therein to be more clearly stated and derived.) In this setting, we immediately have the following connection between $N_{\\mathcal{L}_B}$, the counting function of the reciprocal lengths of $\\mathcal{L}_B$, and the box-counting function $N_B(A,x)$.\n\n\\begin{lemma}\n\\label{lem:CountingFunctionCorrelation}\nFor $x \\in (x_1,\\infty)\\setminus(x_n)_{n \\in \\mathbb{N}}$,\n\\begin{align*}\nN_{\\mathcal{L}_B}(x)=N_B(A,x).\n\\end{align*}\n\\end{lemma}\n\n\\begin{proof}\nThe result follows at once from Definitions \\ref{def:BoxCountingFunction} and \\ref{def:BoxCountingFractalString}.\n\\end{proof}\n\nIn general, Lemma \\ref{lem:CountingFunctionCorrelation} does not hold for $x=x_n$, though equality may hold for certain choices of $N_B(A,x)$. Furthermore, the primary applications of  Lemma \\ref{lem:CountingFunctionCorrelation} are Corollary \\ref{cor:EquivalentZetaFunctions} and Theorem \\ref{thm:JohnnyDarko} where the behavior of $N_B(A,x)$ at $x=x_n$ does not affect the conclusions. Moreover, for a bounded infinite set $A$, the geometric zeta function of the box-counting fractal string $\\mathcal{L}_B$ is given by\n\\begin{align*}\n\\zeta_{\\mathcal{L}_B}(s)&=N_B(A,l_2^{-1})l_1^s+\\sum_{n=2}^{\\infty} (N_B(A,l_{n+1}^{-1})-N_B(A,l_n^{-1})) l_n^s = \\sum_{j=1}^\\infty \\ell_j^s,\n\\end{align*}\nfor $\\textnormal{Re}(s)>D_{\\mathcal{L}_B}$. We take this zeta function to be our \\emph{box-counting zeta function} for a bounded infinite set $A$ in Definition \\ref{def:BoxCountingZetaFunction}. \n\n\n\\begin{definition}\n\\label{def:BoxCountingZetaFunction} \nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$. The \\emph{box-counting zeta function} of $A$, denoted $\\zeta_B$, is the geometric zeta function of the box-counting fractal string $\\mathcal{L}_B$. That is, \n\\begin{align*}\n\\zeta_B(s) := \\zeta_{\\mathcal{L}_B}(s) = \\sum_{n=1}^\\infty m_n l_n^s, \n\\end{align*}\nfor $Re(s) > D_B := D_{\\mathcal{L}_B}$. The (optimum) value $D_B$ is the \\emph{abscissa of convergence} of $\\zeta_B$. The set of \\emph{box-counting complex dimensions} of $A$, denoted $\\calD_B$, is the set of complex dimensions $\\calD_{\\mathcal{L}_B}$ of the box-counting fractal string $\\mathcal{L}_B$.\n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:FiniteBoxCountingFractalString}\nNote that we do not consider the case when $A$ is finite. One may, of course, define the box-counting fractal string $\\mathcal{L}_B$ for such a set as a finite sequence of positive real numbers.  In that case, however, the box-counting zeta function would comprise a finite sum, which would yield an abscissa of convergence $-\\infty$ and no complex dimensions; see Remark \\ref{rmk:AbscissaForFiniteSequences}. That is, in the context of the theory of complex dimensions of fractal strings, the case of finite sets is not very interesting.\n\\end{remark}\n\n\n\\begin{example}[Box-counting zeta function of the Cantor set]\n\\label{eg:CantorSetBoxCountingZetaFunction}\nBy Example \\ref{eg:CantorSetBoxCountingFractalString}, the box-counting fractal string $\\mathcal{L}_B$ of the Cantor set $C$ is given by Equation \\eqref{eqn:CantorSetBoxCountingFractalString}.\nIt follows that for $\\textnormal{Re}(s)>\\log_{3}2$, the box-counting zeta function of $C$ is given by\n\\begin{align*}\n\\zeta_B(s)\t&= 2+\\sum_{n=2}^\\infty 2^{n-1}\\cdot 3^{-(n-1)s} \n= 1+\\frac{1}{1-2\\cdot3^{-s}}. \n\\end{align*}\nThus, $D_B=\\dim_{B}C=\\dim_{M}C=\\log_{3}2$ and $\\zeta_B$ has a meromorphic extension to all of $\\mathbb{C}$ given by the last expression in the above equation. Moreover, we have \n\\begin{align*}\n\\calD_B\t&=\\calD_{CS}=\\calD_{\\mathcal{L}_{CS}}=\\left\\{ \\log_3{2} + i\\frac{2\\pi}{\\log{3}}z : z \\in \\mathbb{Z} \\right\\}.\n\\end{align*}\n\\end{example}\n\n\\begin{example}\n\\label{eg:BoxCountingZetaFunction1DFractal}\nThe box-counting fractal string $\\mathcal{L}_B$ of the 1-dimensional self-similar set $F$ (the attractor of the IFS $\\mathbf{\\Phi}_1$), where $N_B(F,x)$ is as in Definition \\ref{def:BoxCountingFunction}, is given in Example  \\ref{eg:BoxCountingFractalStringOf1DFractal}. Hence, the box-counting zeta function of $F$ is given (for $\\textnormal{Re}(s)>1$) by\n\\begin{align}\n\\label{eqn:1DBoxCountingZetaFunction}\n\\zeta_B(s)\t&= \\left(\\frac{\\sqrt{2}}{2}\\right)^s + \\left(\\left(\\frac{\\sqrt{2}}{2}\\right)^s + \\left(\\frac{\\sqrt{17}}{8}\\right)^s + \\left(\\frac{1}{2}\\right)^s\\right)\n\\sum_{k=0}^\\infty\\left(\\frac{4}{4^s}\\right)^k \\\\\n&= \\left(\\frac{\\sqrt{2}}{2}\\right)^s + \\frac{\\left(\\sqrt{2}/2\\right)^s + \\left(\\sqrt{17}/8\\right)^s + \\left(1/2\\right)^s}{1-4\\cdot 4^{-s}}.\n\\end{align}\nThus, $D_B=\\dim_{B}F=\\dim_{M}F=1$ and $\\zeta_B$ has a meromorphic extension to all of $\\mathbb{C}$ given by the last expression in the above equation. Moreover,  we have \n\\begin{align}\n\\label{eqn:1DBoxCountingComplexDimensions}\n\\calD_B\t&= \\calD_B(\\mathbb{C})=\\left\\{1 + i\\frac{2\\pi}{\\log{4}}z : z \\in \\mathbb{Z} \\right\\}.\n\\end{align}\nNote that the series corresponding to $\\zeta_B(1)$ is divergent. Hence, the fractal string $\\mathcal{L}_B$ does not correspond to an ordinary fractal string (which, by definition, requires $\\zeta_\\mathcal{L}(1)=\\sum_{j=1}^\\infty \\ell_j$ to be convergent). \n\\end{example}\n\n\\begin{remark}\n\\label{rmk:LatticeSelfSimilarIFS}\nThe Cantor set $C$ and the 1-dimensional self-similar set $F$ are each the attractor of a \\emph{lattice} iterated function system; see \\cite[\\S 13.1]{LapvF6} as well as \\cite{LapPe3,LapPe2,LapPeWin}. Essentially, an IFS generated by similarities (i.e., an IFS for which Equation \\eqref{eqn:IFSSelfSimilar} holds) is \\emph{lattice} if, for the distinct values $t_1,\\ldots,t_{N_0}$ among the scaling ratios $r_1,\\ldots,r_N$, there are positive integers $k_j$ where $\\gcd(k_1,\\ldots,k_{N_0})=1$ and a positive real number $0<r<1$ such that $t_j=r^{k_j}$ for each $j=1,\\ldots,N_0$.\\footnote{Equivalently, the precise definition (following \\cite{LapvF6}) is that the \\emph{distinct} scaling ratios generate a multiplicative group of rank 1.} Note that in each case, the box-counting complex dimensions comprise a set of complex numbers with a unique real part (equal to the box-counting dimension) and a vertical (and arithmetic) progression, in both directions, of imaginary parts. \n\nIn the case of the Cantor set $C$, the box-counting complex dimensions $\\calD_B$ coincide with the usual complex dimensions $\\calD_{CS}$. Moreover, the structure of $\\calD_{CS}$ allows for the application of Theorem \\ref{thm:CriterionForMinkowskiMeasurability} and, hence, we conclude (as in \\cite{LapPo1} and \\cite{LapvF6}) that $C$ is not Minkowski measurable.\n\nIn the case of the 1-dimensional self-similar set $F$ of Examples \\ref{eg:BoxCountingFractalStringOf1DFractal} and \\ref{eg:BoxCountingZetaFunction1DFractal}, the set of complex dimensions $\\calD_B$ has no counterpart in the context of usual complex dimensions since $F$ is not the complement of an ordinary fractal string. As such, Theorem \\ref{thm:CriterionForMinkowskiMeasurability} does not apply. Moreover, since $\\dim_MF=1$, the corresponding results in \\cite[\\S 13.1]{LapvF6} do not apply either. \nThis provides motivation for developing a theory of complex dimensions which can take such examples, and many others, into account. The box-counting fractal strings defined in this paper, and investigated further in \\cite{LapRoZu}, provide a first step in developing one such theory. Analogous comments regarding the further development of a higher-dimensional theory of complex dimensions can be made about the results of \\cite{LapRaZu} to be discussed in Section \\ref{sec:DistanceAndTubeZetaFunctions}.\n\\end{remark}\n\nThe next corollary follows readily from Lemma \\ref{lem:CountingFunctionCorrelation} and Proposition \\ref{prop:CountingFunctionDFractalStringD}. It establishes the equivalence of the box-counting zeta function $\\zeta_B$ and an integral transform of the (appropriately truncated) box-counting function $N_B(A,x)$.\n\n\\begin{corollary}\n\\label{cor:EquivalentZetaFunctions}\nLet $A$ be a bounded set. Then \n\\begin{align*}\n\\zeta_B(s) &= \\zeta_{\\mathcal{L}_B}(s) = s\\int_{l_1^{-1}}^{\\infty}x^{-s-1}N_B(A,x)dx,\n\\end{align*}\nfor $Re(s) > D_B$.\n\\end{corollary}\n\nWe close this subsection with a theorem which is a partial statement of our main result, Theorem \\ref{thm:MainResult}. Specifically, the upper box-counting dimension of a bounded infinite set is equal to the abscissa of convergence of the corresponding box-counting zeta function. \n\n\\begin{theorem}\n\\label{thm:JohnnyDarko}\nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$. Then $\\overline{\\dim}_{B}A=D_B$.\n\\end{theorem}\n\n\\begin{proof}\nThe proof follows from a connection made through $D_N$, the asymptotic growth rate of the geometric counting function $N_\\mathcal{L}(x)$ of a fractal string $\\mathcal{L}$ given by Equation \\eqref{eqn:CountingFunctionD}. \n\nLet $\\mathcal{L}=\\mathcal{L}_B$. By Proposition \\ref{prop:CountingFunctionDFractalStringD}, we have $D_\\mathcal{L}=D_B=D_N$. Now, Lemma \\ref{lem:CountingFunctionCorrelation} implies $N_\\mathcal{L}(x)=N_B(A,x)$ for $x \\in (l_1^{-1},\\infty)\\setminus(l_n^{-1})_{n\\in\\mathbb{N}}$. Since these counting functions are nondecreasing, the equation $\\overline{\\dim}_{B}A=D_N$ follows from the formulation  of $\\overline{\\dim}_{B}A$ given in Remark \\ref{rmk:EquivalentUpperBoxDimension}.\n\\end{proof}\n\n\n\n\n\n\\subsection{Tessellation fractal strings and zeta functions} \nIn this subsection, we loosely discuss another type of fractal string defined for a given bounded infinite subset $A$ of $\\mathbb{R}^m$. Unlike the box-counting fractal string $\\mathcal{L}_B$, which is completely determined by the set $A$ and the box-counting function $N_B(A,\\cdot)$, the \\emph{tessellation fractal string} defined here depends on the set $A$, a chosen parameter, and a chosen family of tessellations of $\\mathbb{R}^m$.\n\nFirst, choose a scaling parameter $\\lambda\\in(0,1)$. For any $n\\in\\mathbb{N}$, consider the $n$-th tessellation of $\\mathbb{R}^m$ defined by the family of cubes of length $\\lambda^n$ (obtained by taking translates of the cube $[0,\\lambda^n]^m$ in $\\mathbb{R}^m$). Henceforth, the number of cubes of the $n$-th tessellation that intersect $A$ is denoted by $m_n(\\lambda)$. Let the scale $l_n(\\lambda):=\\lambda^n$ be of multiplicity $m_n(\\lambda)$. This defines the box-counting fractal string $\\mathcal{L}(A,\\lambda)=(\\ell_j)_{j\\in\\mathbb{N}}$, where $(\\ell_j)_{j\\in\\mathbb{N}}$ is the sequence starting with $l_1$ with multiplicity $m_1$, $l_2$ with multiplicity $m_2$, and so on. The geometric counting function $N_{\\mathcal{L}(A,\\lambda)}(x)=\\#\\{j\\in\\mathbb{N}: \\ell_j^{-1}\\ge x\\}$ of the fractal string is then well defined. \n\nMore generally, let $U$ be a compact subset of $\\mathbb{R}^m$ with nonempty interior, satisfying the following properties:\n\\begin{enumerate}\n\\addtolength{\\itemsep}{0.3\\baselineskip}\n\\item[(a)] there exists a countable family of isometric maps $f_j:\\mathbb{R}^m\\to\\mathbb{R}^m$ such that the the family of sets $V_j=f_j(U)$, $j\\ge1$, is a cover of $\\mathbb{R}^m$;\n\\item[(b)] for $j\\neq k$, the interiors of $V_j$ and $V_k$ are disjoint.\n\\end{enumerate}\nWe say that the family $(V_j)_{j\\ge1}$ is a \\emph{tessellation of $\\mathbb{R}^m$}, generated by the basic shape (or `tile') $U$ and the family of isometries. Also, we say in short that \\emph{$U$ tessellates~$\\mathbb{R}^m$}. Note that if $\\lambda\\in(0,1)$ is a fixed real number, then the basic shape generates a sequence of tessellations indexed by $n\\in\\mathbb{N}$: $(\\lambda^n V_j)_{j\\ge1}$. The family $(\\lambda^n V_j)_{j\\ge1}$ is called the \\emph{$n$-th tessellation of $\\mathbb{R}^m$}, generated by the basic shape (or tile) $U$, the family of isometries $(f_j)_{j\\ge1}$, and $\\lambda\\in(0,1)$.\n\nLet $A$ be a given bounded set in $\\mathbb{R}^m$. Define $m_n(\\lambda)=m_n(A,U,\\lambda)$ analogously as above, by counting the number of elements of the $n$-th tessellation which intersect $A$. The \\emph{tessellation fractal string} $\\mathcal{L}(A,U,\\lambda)$ of the set $A$ is then  the fractal string defined by\n\\begin{align*}\n\\mathcal{L}(A,U,\\lambda)\t&:=\\{l_n(\\lambda)=\\lambda^n:\\mbox{ $l_n(\\lambda)$ has multiplicity $m_n(\\lambda)$, $n\\in\\mathbb{N}$}\\}=(\\ell_j)_{j\\in\\mathbb{N}}.\n\\end{align*}\nThe middle set is in fact a \\emph{multiset}, by which we mean that its elements repeat with prescribed multiplicity. Using arguments analogous to those from \\cite[pp.\\ 38--39]{Falc}, we obtain that\n\\begin{equation}\n\\label{dimBA}\n\\overline\\dim_BA=\\limsup_{n\\to\\infty}\\frac{\\log m_n(\\lambda)}{\\log\\lambda^{-n}}.\n\\end{equation}\n(Also, see \\cite[p.\\ 41]{Falc} or \\cite[p.\\ 24]{Tri}.) Here,  we have also used a version of Proposition~\\ref{prop:GeometricSequenceSufficient} above.\n\nThe geometric zeta function of the tessellation fractal string $\\mathcal{L}(A,U,\\lambda)$, called the \\emph{tessellation zeta function}, is given by \n\\begin{align}\n\\label{zeta_tess}\n\\zeta_{\\mathcal{L}(A,U,\\lambda)}(s)\t&=\\sum_{j=1}^{\\infty}\\ell_j^s=\\sum_{n=1}^{\\infty} m_n(\\lambda)\\,\\lambda^{ns}\n\\end{align}\nfor $\\textnormal{Re}(s)$ large enough. Also, when defined accordingly, the set of complex dimensions $\\calD_{\\mathcal{L}(A,U,\\lambda)}$ of the tessellation fractal string $\\mathcal{L}(A,U,\\lambda)$ is called the set of \\emph{tessellation complex dimensions} of $A$ (relative to the tessellation associated with $U$ and $\\lambda$).\n\nThe main result regarding this zeta function is the following theorem.\n\n\\begin{theorem}\n\\label{thm:TessellationDimension}\nLet $U \\subset \\mathbb{R}^m$ be a compact set with nonempty interior, which tessellates $\\mathbb{R}^m$. Then, the upper box-counting dimension of a bounded infinite set $A$ in $\\mathbb{R}^m$ is equal to the abscissa of convergence $D_{\\mathcal{L}(A,U,\\lambda)}$ of the geometric zeta function \\eqref{zeta_tess} of its tessellation fractal string. That is, \n\\begin{align*}\n\\overline{\\dim}_{B}A=D_{\\mathcal{L}(A,U,\\lambda)}.\n\\end{align*}\n\\end{theorem}\n\n\n\\begin{proof}\nUsing Cauchy's criterion for convergence, we obtain that the series \\eqref{zeta_tess} converges for all $s\\in\\mathbb{C}$ such that\n$$\n\\limsup_{n\\to\\infty}m_n(\\lambda)^{1/n}\\lambda^{\\operatorname{Re}(s)}<1,\n$$\nthat is,\n$$\n\\operatorname{Re}(s)>\\frac{\\log(\\limsup_{n\\to\\infty}m_n(\\lambda)^{1/n})}{\\log\\lambda^{-1}}.\n$$\nThe series (\\ref{zeta_tess}) diverges if we have the opposite inequality. Therefore, the abscissa of convergence of (\\ref{zeta_tess}) is\n\\begin{equation}\n\\label{dimD}\nD_{\\mathcal{L}(A,U,\\lambda)}=\\frac{\\log(\\limsup_{n\\to\\infty}m_n(\\lambda)^{1/n})}{\\log\\lambda^{-1}}=\\limsup_{n\\to\\infty}\\frac{\\log m_n(\\lambda)}{\\log\\lambda^{-n}}.\n\\end{equation}\nIn light of \\eqref{dimBA} and \\eqref{dimD}, we deduce that $\\overline\\dim_BA=D_{\\mathcal{L}(A,U,\\lambda)}$.\n\\end{proof}\n\n\\begin{remark}\nThe value of $D_{\\mathcal{L}(A,U,\\lambda)}$ is independent of the choice of $\\lambda\\in(0,1)$, since by Theorem~\\ref{thm:TessellationDimension} its\nvalue is equal to $\\overline\\dim_BA$. In concrete applications we choose the basic shape $U$\nand $\\lambda\\in(0,1)$ that are best suited to the geometry of $A$. For example, if $A$ is the triadic Cantor set, we take $U=[0,1]$ and $\\lambda=1/3$, while for the\nSierpinski gasket we take $U$ to be an equilateral triangle and $\\lambda=1/2$.\n\\end{remark}\n\n\n\\begin{example} \nLet $F$ be the 1-dimensional self-similar set from Examples \\ref{eg:BoxCountingFractalStringOf1DFractal} and \\ref{eg:BoxCountingZetaFunction1DFractal}. We define $U$ as the unit square $[0,1]^2$ and $\\lambda=1/4$. Here, the scale $l_n(1/4)=1/4^n$ occurs with multiplicity $m_n(1/4)=9\\cdot4^n$, defining the corresponding tessellation fractal string $\\mathcal{L}(F,U,1/4)$. For $\\textnormal{Re}(s)>1$, the tessellation zeta function is given by \n$$\n\\zeta_{\\mathcal{L}(F,U,1/4)}(s) = \\sum_{n=1}^\\infty 9\\cdot4^n\\cdot 4^{-ns}=\\frac9{4^{s-1}-1}.\n$$\nAccording to Theorem~\\ref{thm:TessellationDimension}, the abscissa of convergence  $D_{\\mathcal{L}(F,U,1/4)}=1$ equal to the box-counting dimension of $F$. It follows that $\\zeta_{\\mathcal{L}(F,U,1/4)}(s)$ has a meromorphic extension to all of $\\mathbb{C}$ given by $9(4^{s-1}-1)^{-1}$. Furthermore, the set of tessellation complex dimensions is equal to the set $\\calD_B$ of box-counting complex dimensions given in~ \\eqref{eqn:1DBoxCountingComplexDimensions}. That is,\n\\[\n\\calD_{\\mathcal{L}(F,U,1/4)}=\\calD_B=\\left\\{ 1 + i\\frac{2\\pi}{\\log 4}z : z\\in\\mathbb{Z} \\right\\}.\n\\]\nNote that the tessellation fractal string $\\mathcal{L}(F,U,1/4)$ is unbounded, in the sense that the series given by $\\zeta_{\\mathcal{L}(F,U,1/4)}(1)$ is divergent.\n\nAnalogous results hold regarding the Cantor set $C$ and its (classical and box-counting) fractal strings, zeta functions, and complex dimensions. Further (higher-dimensional) examples will be studied in \\cite{LapRoZu}.\n\\end{example}\n\n\n\n\n\n\n\\section{Distance and tube zeta functions}\n\\label{sec:DistanceAndTubeZetaFunctions}\n\nIn this section, we deal with a class of zeta functions introduced by the first author during the 2009 ISAAC Conference at the University of Catania in Sicily, Italy. More generally, the main results of this section are obtained in the forthcoming paper \\cite{LapRaZu}, written by the first and third authors, along with Goran Radunovi\\'c. We state here only some of the basic results, without attempting to work at the greatest level of generality. We refer to \\cite{LapRaZu} for more general statements and additional results and illustrative examples.\n\nThe following definition can be found in \\cite{LapRaZu}. \n\\begin{definition} \n\\label{def:DistanceZetaFunction}\nLet $A \\subset \\mathbb{R}^m$ be bounded. The \\emph{distance zeta function} of $A$, denoted $\\zeta_d$, is defined by\n\\begin{align}\n\\label{eqn:DistanceZetaFunction}\n\\zeta_d(s)\t&:=\\int_{A_\\varepsilon}d(x,A)^{s-m}dx\n\\end{align}\nfor $\\textnormal{Re}(s)>D_d$, where $D_d=D_d(A)$ denotes the abscissa of convergence of the distance zeta function $\\zeta_d$ and $\\varepsilon$ is a fixed positive number. \n\\end{definition}\n\n\\begin{remark}\n\\label{rmk:IndependentOfVarepsilon}\nIt is shown in \\cite{LapRaZu} that \nchanging the value of $\\varepsilon$ modifies the distance zeta function by adding an entire function to $\\zeta_d$. Hence, the main properties of $\\zeta_d$ do not depend on the choice of $\\varepsilon>0$. As a result, this is also the case for $D_d$, the abscissa of convergence of $\\zeta_d$ (cf. Theorem \\ref{thm:DistanceZetaFunctionUpperBoxDimension}), and $\\textnormal{res}(\\zeta_d;D_d)$, the residue of $\\zeta_d$ at $s=D_d$ (cf. Theorem \\ref{thm:DistanceZetaFunctionAndMinkowskiContent}).\n\\end{remark}\n\nThe distance zeta function can be used as an effective tool in the computation of the box-counting dimensions of various subsets $A$ of some Euclidean space; see \\cite{LapRaZu}. Indeed, one of the basic results concerning the distance zeta function is given in the following theorem, which is Theorem 1 in \\cite{LapRaZu}. Note: unlike in Theorem \\ref{thm:JohnnyDarko} above, we allow $A$ to be finite here.\n\n\\begin{theorem}\n\\label{thm:DistanceZetaFunctionUpperBoxDimension}\nLet $A$ be a nonempty bounded subset of $\\mathbb{R}^m$. Then $D_d=\\overline{\\dim}_{B}A$.\n\\end{theorem}\n\n\\begin{corollary}\n\\label{cor:holomorphic}\n$\\zeta_d$ is holomorphic in the half-plane $\\textnormal{Re}(s)>\\overline{\\dim}_BA$. Furthermore, this open right half-plane is the largest one in which $\\zeta_d(s)$ is holomorphic.  \n\\end{corollary}\n\n\\begin{remark}\n\\label{rmk:LowerBoxCountingDimensionFromDistanceZetaFunction}\nWe do not know if the value of the \\emph{lower} box-counting dimension $\\underline{\\dim}_BA$ can be computed from the distance zeta function $\\zeta_d$.\n\\end{remark}\n\nIt is shown in \\cite{LapRaZu} that the distance zeta function represents a natural extension of the geometric zeta function $\\zeta_{\\mathcal{L}}$ of a bounded (i.e., summable) fractal string $\\mathcal{L}=(\\ell_j)_{j\\in\\mathbb{N}}$. Indeed, we can identify the string with an ordinary fractal string of the form $\\Omega=\\cup_{j=1}^\\infty I_j$, where $I_j:=(a_{j+1},a_j)$ and $a_j:=\\sum_{k\\ge j}\\ell_k$. Note that $|I_j|=\\ell_j$. Defining $A=\\{a_j\\}_{j=1}^\\infty\\subset\\mathbb{R}$, it is easy to see that $\\zeta_d(s)=a(s)\\zeta_{\\mathcal{L}}(s)+b(s)$, where $a(s)$ vanishes nowhere and $a(s)$ and $b(s)$ are explicit meromorphic functions in the complex plane with (typically) a pole at the origin. Hence, the zeta functions  $\\zeta_\\mathcal{L}$ and $\\zeta_d$ have the same abscissa of convergence. \\emph{It follows that if $\\mathcal{L}$ is nontrivial} (\\emph{i.e., has infinitely many lengths}), \\emph{then}\\footnote{If we allow $\\mathcal{L}$ to be trivial, then one should replace $D_d(A)$ and $D_\\mathcal{L}$ with $\\max\\{D_d(A),0\\}$ and $\\max\\{D_\\mathcal{L},0\\}$, respectively, in Equation \\eqref{eqn:DistanceDimensionEquivalence}.}\n\\begin{align}\n\\label{eqn:DistanceDimensionEquivalence}\nD_d(A)=D_\\mathcal{L}=\\overline{\\dim}_BA=\\overline{\\dim}_B(\\partial\\Omega).\n\\end{align}\n\n\n\\subsection{Minkowski content and residue of the distance zeta function}\n\\label{sec:MinkowskiContentAndResidueOfTheDistanceZetaFunction}\n\nA remarkable property of the distance zeta function is that its residue at $s=D_d$ is closely related to the $D_d$-dimensional Minkowski content of $A$; see \\cite{LapRaZu}.\n\n\\begin{theorem}\n\\label{thm:DistanceZetaFunctionAndMinkowskiContent}\nLet $A$ be a nonempty bounded set in $\\mathbb{R}^m$.\nAssuming that the distance zeta function can be meromorphically extended to a neighborhood of $s=D_d$ and $D_d<m$, then for its residue at $s=D_d$ we have that\n\\begin{align}\n\\label{inequality}\n(m-D_d){\\mathscr{M}}_*^{D_d}\\le \\operatorname{res}(\\zeta_d(s);D_d)\\le(m-D_d){\\mathscr{M}}^{*D_d}.\n\\end{align}\nIf, in addition, $A$ is Minkowski measurable, it then follows that\n\\begin{align}\n\\label{eqn:MinkowskiContentAsResidue}\n\\operatorname{res}(\\zeta_d(s);D_d)=(m-D_d){\\mathscr{M}}^{D_d}.\n\\end{align}\n\\end{theorem}\n\nThe last part of this result (namely, Equation \\eqref{eqn:MinkowskiContentAsResidue}) generalizes the corresponding one obtained in \\cite{LapvF6} in the context of ordinary fractal strings to the case of bounded sets in Euclidean spaces; see \\cite{LapRaZu}.\n\n\\begin{example}\nIt can be shown that, in the case of the Cantor set $C$, we have strict inequalities in Equation \\eqref{inequality}. Indeed, in this case $m=1$, $D_d=\\log_{3}2$, and\n$$\n\\operatorname{res}(\\zeta_d(s);D_d)=\\frac1{\\log2}\\,2^{-D_d},\n$$\nwhereas the values of the lower and upper $D_d$-dimensional Minkowski contents have been computed in \\cite[Theorem 2.16]{LapvF6} (as well as earlier in \\cite{LapPo1}):\n$$\n{\\mathscr{M}}_*^{D_d}(A)=\\frac1{D_d}\\left(\\frac{2D_d}{1-D_d} \\right)^{1-D_d} ,\\quad\n{\\mathscr{M}}^{*D_d}(A)=2^{2-D_d}.\n$$\nThis is a special case of an example in~\\cite{LapRaZu} dealing with generalized Cantor sets. Generalized Cantor strings, which are a certain type of generalized fractal strings, and their (geometric and spectral) oscillations are studied in detail in \\cite[Ch.~10]{LapvF6}.\n\\end{example}\n\n\\begin{remark}\nAn open problem is to determine whether there exists a set $A$ such that one of the inequalities in Equation \\eqref{inequality} is strict and the other is an equality.\n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:Resman}\nAccording to a recent result due to Maja Resman in \\cite{Res}, we know that if $A$ is Minkowski measurable, then the value of the \\emph{normalized} $D_d$-\\emph{dimensional Minkowski content} of a bounded set $A \\subset \\mathbb{R}^m$,\\footnote{This choice of normalized Minkowski content is well known in the literature; see, e.g., \\cite{Fdrr}.} defined by\n\\begin{align}\n\\label{eqn:DdMinkowskiContent}\n\\frac{{\\mathscr{M}}^{D_d}(A)}{\\omega(m-D_d)},\n\\end{align}\nis independent of the ambient dimension $m$. Here, for $t>0$, we let \n\\[\n\\omega(t):=2\\pi^{t/2}t^{-1}\\Gamma(t/2)^{-1},\n\\]\nwhere $\\Gamma$ is the classic Gamma function. For any positive integer $k$, $\\omega(k)$ is equal to the $k$-dimensional Lebesgue measure of the unit ball in $\\mathbb{R}^k$. In other words, the value given in Equation \\eqref{eqn:DdMinkowskiContent} is intrinsic to the set $A$ and hence independent of the embedding of $A$ in $\\mathbb{R}^k$. Therefore, we may ask if the value of the normalized residue,\n\\begin{align*}\n\\frac{\\mbox{res}(\\zeta_d(s);D_d)}{(m-D_d)\\,\\omega(m-D_d)},\n\\end{align*}\nis also independent of $m$. Combining the preceding two results (namely, Theorems \\ref{thm:DistanceZetaFunctionUpperBoxDimension} and \\ref{thm:DistanceZetaFunctionAndMinkowskiContent}), we immediately deduce that if $A$ is Minkowski measurable, then the answer is positive.\n\\end{remark}\n\n\\subsection{Tube zeta function}\n\\label{sec:TubeZetaFunction}\n\nGiven $\\varepsilon>0$, it is also natural to introduce the following zeta function of a bounded set $A$ in $\\mathbb{R}^m$, involving the tube around $A$ (which we view as the mapping $t\\mapsto |A_t|$, for $0\\leq t\\leq\\varepsilon$):\\footnote{In the sequel, $|A_t|=|A_t|_m$ denotes the $m$-dimensional volume (Lebesgue measure) of $A_t$, the $t$-neighborhood of $A\\subset\\mathbb{R}^m$. In our earlier notation, we have $|A_t|=\\textnormal{vol}_m(A_t)$.}\n\\begin{align}\n\\label{tube_zeta}\n\\tilde\\zeta_A(s)\t&=\\int_0^\\varepsilon t^{s-m-1}|A_t|\\,dt,\n\\end{align}\nfor $\\textnormal{Re}(s)$ sufficiently large, where $\\varepsilon$ is a fixed positive number. Hence, $\\tilde\\zeta_A$ is called the \\emph{tube zeta function} of $A$. Its abscissa of convergence is equal to $\\overline{\\dim}_{B}A$, which follows immediately from Theorems \\ref{thm:DistanceZetaFunctionUpperBoxDimension} above and \\ref{thm:DistanceAndTubeZetaFunctions} below. Tube zeta functions are closely related to distance zeta functions, as shown by the following result; see \\cite{LapRaZu}.\n\n\\begin{theorem}\n\\label{thm:DistanceAndTubeZetaFunctions}\nIf $A\\subset\\mathbb{R}^m$ and $\\operatorname{Re}(s)>\\overline{\\dim}_BA$, then for any $\\varepsilon>0$,\n\\begin{equation}\n\\label{identity}\n\\zeta_d(s)\t=\\varepsilon^{s-m}|A_\\varepsilon|+(m-s)\\tilde\\zeta_A(s).\n\\end{equation}\n\\end{theorem}\n\nThe proof of this result when $s\\in\\mathbb{R}$ follows, for example, from \\cite[Theorem~2.9(a)]{Zu0}, and the proof of the cited theorem is based on integration by parts. By analytic continuation and in light of Corollary \\ref{cor:holomorphic}, the identity \\eqref{identity} is then extended to complex values of $s$ such that $\\textnormal{Re}(s)>\\overline{\\dim}_BA$.\n\nIt follows from \\eqref{identity} that the abscissae of convergence of the zeta functions $\\zeta_d$ and $\\tilde\\zeta_A$ are the same, and therefore also coincide with $\\overline{\\dim}_BA$. This identity, \\eqref{identity}, extends to the  $m$-dimensional case \\cite[identity (13.129) in Lemma 13.110, p.~442]{LapRaZu}, which has been formulated in the context of $p$-adic fractal strings and ordinary (real) fractal strings. (See also \\cite{LapLuvF1}.) Using \\eqref{identity} and Theorem \\ref{thm:DistanceZetaFunctionAndMinkowskiContent}, it is easy to derive the following consequence; see \\cite{LapRaZu}.\n\n\\begin{corollary} \n\\label{cor_res}\nIf $D=\\dim_BA$ exists, $D<m$, and there exists a meromorphic extension of $\\tilde\\zeta_A(s)$ to a neighborhood of $s=D$, then\n\\begin{align*}\n&{\\mathscr{M}}_*^D \\le\\operatorname{res}(\\tilde\\zeta_A(s);D)\\le {\\mathscr{M}}^{*D}.\n\\end{align*}\nIn particular, if $A$ is Minkowski measurable, then \n\\begin{align*}\n\\operatorname{res}(\\tilde\\zeta_A(s);D)\t&={\\mathscr{M}}^D.\n\\end{align*}\n\\end{corollary}\n\nAs we can see, the tube zeta function is ideally suited to study the Minkowski content.\n\nIn Corollary~\\ref{cor_res}, we have assumed the existence of a meromorphic extension of the tube zeta function to a neighborhood of $s=D$. This condition can be ensured under fairly general conditions. We provide a result from~\\cite{LapRaZu} dealing with the case of Minkowski measurable sets. Non-Minkowski measurable sets can be treated as well; see \\cite{LapRaZu}, along with Remark \\ref{rmk:NonMinkowskiAnalogs} and Theorem \\ref{thm:NonMinkowskiMeasurableCase}. \n\n\\begin{theorem}[Minkowski measurable case]\n\\label{mer}\nLet $A$ be a subset of $\\mathbb{R}^m$ such that there exist  $\\alpha>0$, $M\\in(0,\\infty)$ and $D\\in[0,m]$,  satisfying\n\\begin{align}\n\\label{eqn:MinkowskiMeasurableCase}\n|A_t| &= t^{m-D}\\left(M+O(t^\\alpha)\\right)\\quad\\mbox{as $t\\to 0^+$.}\n\\end{align}\nThen $A$ is Minkowski measurable, $\\operatorname{dim}_BA=D$, and $\\mathscr{M}^D(A)=M$. Furthermore, the tube zeta function $\\tilde{\\zeta}_A(s)$ has for abscissa of convergence  $D(\\tilde\\zeta_A)=D$, and it admits a \\emph{(}necessarily unique\\emph{)} meromorphic continuation {\\rm({\\it at least})} to the half-plane $\\{\\operatorname{Re}(s)>D-\\alpha\\}$. The only pole in this half-plane is $s=D$; it is simple, and $\\operatorname{res}(\\tilde\\zeta_A;D)=M$.\n\\end{theorem}\n\nAn analogous result holds also for the distance function of $A$. Theorem~\\ref{mer} shows that  the relevant information concerning the possible existence of a nontrivial meromorphic extension of the tube (or the distance) zeta function associated with~$A$, is encoded in the second term of the asymptotic expansion (as $t \\to 0^+$) of the tube function $t\\mapsto|A_t|$. Various extensions of this result and examples can be found in~\\cite{LapRaZu}, as we next discuss. For example, in Theorem \\ref{mer}, the conclusion for $\\zeta_d$ would be the same except for the fact that $\\operatorname{res}(\\zeta_d(s);D)=(m-D)M$ (and $D=D_d(A)$, in our earlier notation from Equation \\eqref{eqn:DistanceDimensionEquivalence}, for instance). \n\n\\begin{remark}\n\\label{rmk:NonMinkowskiAnalogs}\nIn \\cite{LapRaZu}, one can find suitable analogs of Theorem \\ref{mer} for non-Minkowski measurable sets, both in the case where the underlying scaling behavior of $|A_t|$ is log-periodic (as for the Cantor set or the Sierpinski gasket and carpet, for example) and in more general, non-periodic situations. Furthermore, in that case, the (visible) complex dimensions of $A$ (i.e., the poles of $\\tilde{\\zeta}_A$ in the half-plane $\\{\\textnormal{Re}(s)>D-\\alpha\\}$, with $\\alpha>0$, are also determined, and shown to consist of a vertical infinite arithmetic progression located on the `critical line' $\\{\\textnormal{Re}(s)=D\\}$ (much as for the Cantor string and other lattice self-similar strings); see Theorem \\ref{thm:NonMinkowskiMeasurableCase} below for a typical sample theorem. In \\cite{LapRaZu}, the case of so-called `quasi-periodic' fractals is also considered in this and related contexts. In the latter situation, one appeals, in particular, to well-known (and rather sophisticated) number theoretic theorems asserting the transcendentality (and hence, the irrationality) of certain expressions. It is noteworthy that (in light of Theorem \\ref{thm:DistanceAndTubeZetaFunctions}) all of these results concerning $\\tilde{\\zeta}_A$ have precise counterparts for the distance zeta function $\\zeta_d$; see \\cite{LapRaZu}.  \n\\end{remark}\n\nIn order to state more precisely a sample theorem in the non-Minkowski measurable case, we introduce the following hypothesis (LP) (log-periodic) and notation:\n\n\\begin{itemize} \n\\item[(LP)] Let $A\\subset\\mathbb{R}^m$ be a bounded set such that there exists $D\\geq 0, \\alpha>0$ and a periodic function $G:\\mathbb{R}\\to [0,\\infty)$ with minimal period $T>0$, satisfying\n\\end{itemize}\n\\begin{align}\n\\label{eqn:NonConstantPeriodic}\n|A_t| &= t^{m-D}(G(\\log{t^{-1}})+O(t^\\alpha)), \\quad \\textnormal{as } t\\to 0^+.\n\\end{align}\nIn the sequel, let\n\\begin{align}\n\\label{eqn:NonConstantPeriodicIntegral}\n\\hat{G}_0(t) &= \\int_0^T e^{-2\\pi it\\tau}G(\\tau)\\,d\\tau, \\quad \\textnormal{for } t\\in\\mathbb{R}.\n\\end{align}\nNote that $G$ is nonconstant since we have assumed it to have a positive minimal period, and that $\\hat{G}_0$ is (essentially) the Fourier transform of the cut-off function $G_0$ of $G$ to $[0,T]$.\n\n\\begin{theorem}[Non-Minkowski measurable case]\n\\label{thm:NonMinkowskiMeasurableCase}\nAssume that the bounded set $A\\subset\\mathbb{R}^m$ satisfies the log-periodicity property \\emph{(LP)} above.  Then $\\dim_BA$ exists, $\\dim_BA=D$, $G$ is continuous and\\footnote{It follows that the restriction of $G$ to $[0,T]$ has for range $[\\calM^D_*(A),\\calM^{*D}(A)]$.}\n\\begin{align}\n\\label{eqn:UpperLowerMinkowskiContentNonMinkCase}\n\\calM^D_*(A)&=\\min{G}, \\quad \\calM^{*D}(A)=\\max{G}.\n\\end{align}\nHence, $A$ is not Minkowski measurable and \\emph{(}provided $G>0$\\emph{)} is non-degenerate, i.e., $0<\\calM^D_*(A)<\\calM^{*D}(A)<\\infty$.\n\nFurthermore, $\\tilde{\\zeta}_A$ admits a \\emph{(}necessarily unique\\emph{)} meromorphic continuation to the half-plane $\\{\\textnormal{Re}(s)>D-\\alpha\\}$ with only poles \n\\begin{align}\n\\label{eqn:PolesNonMinkowskiCase}\n\\mathcal{P}(\\tilde{\\zeta}_A)\t&=\\left\\lbrace s_k:=D+\\frac{2\\pi}{T}ik: \\hat{G}_0\\left(\\frac{k}{T}\\right)\\neq 0, k\\in\\mathbb{Z}\\right\\rbrace,\n\\end{align}\nwhere $\\hat{G}_0$ is given by \\eqref{eqn:NonConstantPeriodicIntegral}. These poles are all simple and\\footnote{In addition, $|\\textnormal{res}(\\tilde{\\zeta}_A;s_k)|\\leq\\frac{1}{T}\\int_0^TG(\\tau)\\,d\\tau$ for all $k\\in\\mathbb{Z}$, and $\\textnormal{res}(\\tilde{\\zeta}_A;s_k)\\to 0$ as $|k|\\to\\infty$.}\n\\begin{align}\n\\label{eqn:ResiduesNonMinkCase}\n\\textnormal{res}(\\tilde{\\zeta}_A;s_k)\t&= \\frac{1}{T}\\hat{G}_0\\left(\\frac{k}{T}\\right), \\quad \\textnormal{for all }k\\in\\mathbb{Z}.\n\\end{align}\nIn particular, for $k=0$, we have that \n\\begin{align}\n\\label{eqn:ResidueAsIntegralNonMinkCase}\n\\textnormal{res}(\\tilde{\\zeta}_A;D)\t&=\\frac{1}{T}\\int_0^TG(\\tau)\\,d\\tau.\n\\end{align}\n\nFinally, $A$ admits an average Minkowski content $\\calM^D_{av}$ \\emph{(}which lies in $(0,\\infty)$ if $G>0$\\emph{)} that is also given by \\eqref{eqn:ResidueAsIntegralNonMinkCase}.\\footnote{This average Minkowski content is defined exactly as in \\cite[Definition~8.29,\\S 8.4.3]{LapvF6}, except for the fact that $1-D$ is replaced with $m-D$.}   \n\\end{theorem} \n\n\\begin{remark}\n\\label{rmk:CounterpartTheoremForDistanceZetaFunctions}\nAs was alluded to in Remark \\ref{rmk:NonMinkowskiAnalogs}, Theorem \\ref{thm:NonMinkowskiMeasurableCase} has a precise counterpart for the distance zeta function $\\zeta_d$. In fact, under the same assumption (LP), the same conclusions as in Theorem \\ref{thm:NonMinkowskiMeasurableCase} hold, except for the fact that the counterpart of \\eqref{eqn:ResiduesNonMinkCase} and \\eqref{eqn:ResidueAsIntegralNonMinkCase}, respectively, reads as follows (with $D=D_{d}(A)$, as in Equation \\eqref{eqn:DistanceDimensionEquivalence} and Theorems \\ref{thm:DistanceZetaFunctionAndMinkowskiContent} and \\ref{thm:NonMinkowskiMeasurableCase}):\n\\begin{align}\n\\label{eqn:ResiduesNonMinkCaseDistance}\n\\textnormal{res}(\\zeta_d;s_k)\t&= \\frac{m-s_k}{T}\\hat{G}_0\\left(\\frac{k}{T}\\right), \\quad \\textnormal{for all }k\\in\\mathbb{Z},\n\\end{align}\nand \n\\begin{align}\n\\label{eqn:ResidueAsIntegralNonMinkCaseDistance}\n\\textnormal{res}(\\zeta_d;D)\t&=(m-D)\\frac{1}{T}\\int_0^TG(\\tau)\\,d\\tau,\n\\end{align}\nso that $\\calM^D_{av}(A)$, the average Minkowski content of $A$, is now given by \n\\begin{align}\n\\label{eqn:AverageMinkowskiContent}\n\\calM^D_{av}(A)\t&=\\frac{1}{T}\\int_0^TG(\\tau)\\,d\\tau=\\frac{1}{m-D}\\textnormal{res}(\\zeta_d;D).\n\\end{align}\n\\end{remark}\n\n\\begin{example}\n\\label{eg:IllustratingNonMinkCase}\nTheorem \\ref{thm:NonMinkowskiMeasurableCase} (and its counterpart for $\\zeta_d$ discussed in Remark \\ref{rmk:CounterpartTheoremForDistanceZetaFunctions}) can be illustrated, for instance, by the Cantor set or string (see \\cite{LapRaZu} and compare with \\cite[\\S 1.1.2]{LapvF6}, including Figures 1.4 and 1.5) and the Sierpinski carpet (as well as, similarly, by the Sierpinski gasket). For the Sierpinski carpet $A$, as discussed in \\cite{LapRaZu}, we have $D=\\log_{3}8$, $\\alpha=D-1$, and $T=\\log{3}$. Hence, both $\\tilde{\\zeta}_A$ and $\\zeta_d$ have a meromorphic continuation to $\\left\\lbrace \\textnormal{Re}(s)>1\\right\\rbrace$, with set of (simple) poles (the visible complex dimensions of $A$) given by\n\\begin{align}\n\\label{eqn:IllustrationNonMinkCaseComplexDimensions}\n\\mathcal{P}(\\tilde{\\zeta}_A)\t&= \\mathcal{P}(\\zeta_d) = \\left\\lbrace D+\\frac{2\\pi}{T}ik: \\hat{G}_0\\left(\\frac{k}{T}\\right)\\neq 0, k\\in\\mathbb{Z}\\right\\rbrace.\n\\end{align}\nActually, in this case, both $\\tilde{\\zeta}_A$ and $\\zeta_d$ can be meromorphically extended to all of $\\mathbb{C}$.\n\\end{example}\n\nA number of other results concerning the existence of meromorphic continuation of $\\tilde{\\zeta}_A$ and $\\zeta_d$ (as well as $\\zeta_\\mathcal{L}$, in the case of fractal strings) and the resulting structure of the poles (the visible complex dimensions) can be found in \\cite{LapRaZu}, under various assumptions on the bounded set $A\\subset\\mathbb{R}^m$ (or on the fractal string $\\mathcal{L}$). Moreover, a number of other applications of the distance and tube zeta functions are provided in \\cite{LapRaZu}, in order to study various classes of fractals, including fractal chirps and `zigzagging' fractals.\n\n\\begin{remark}\nThe box-counting zeta function $\\zeta_B$ of a set $A \\subset \\mathbb{R}^m$ given by Definition \\ref{def:BoxCountingZetaFunction} is closely related to the tube zeta function $\\tilde\\zeta_A$. To see this, it suffices to perform the change of variables $x=t^{-1}$ in Equation \\eqref{tube_zeta} and compare with Corollary~\\ref{cor:EquivalentZetaFunctions}. Note that for $x>0$, we have  (under suitable hypotheses) $|A_{1/x}|\\asymp x^{-m}N_B(A,x)$ as $x\\to\\infty$. Here, $N_B(A,x)$ is defined as the number of $x^{-1}$-mesh cubes that intersect $A$; see (iv) in Remark \\ref{rmk:VariousCountingFunctions}. It is clear, however, that these two zeta functions are in general not equal to each other. Moreover, we do not know if the corresponding two sets of complex dimensions of $A$, associated with these two zeta functions, coincide.\n\\end{remark}\n\nVarious generalizations of the notion of distance zeta function are possible. One of them, which is especially interesting, deals with zeta functions associated to relative fractal drums. By a \\emph{relative fractal drum}, introduced in \\cite{LapRaZu}, we mean an ordered pair $(A,\\Omega)$, where $A$ is an arbitrary nonempty subset of $\\mathbb{R}^m$, and $\\Omega$ is an open subset such that $A_\\varepsilon$ contains $\\Omega$ for some positive $\\varepsilon$, and the $m$-dimensional Lebesgue measure of $\\Omega$ is finite. (Note that both $A$ and $\\Omega$ are now allowed to be unbounded.) The corresponding \\emph{relative zeta function} (or the distance zeta function of the relative fractal drum), also introduced in \\cite{LapRaZu}, is defined in much the same way as in Equation \\eqref{eqn:DistanceZetaFunction}:\n\\begin{align*}\n\\zeta_d(s;A,\\Omega)\t&:=\\int_{\\Omega}d(x,A)^{s-m}dx.\n\\end{align*}\n\nIt is possible to show that the abscissa of convergence of the relative zeta function is equal to the relative box dimension $\\overline\\dim_B(A,\\Omega)$; see \\cite{LapRaZu} for details and illustrative examples.\\footnote{We caution the reader that, in general, the relative upper box dimension (defined as an infimum over $\\alpha \\in \\mathbb{R}$ instead of over $\\alpha\\geq 0$) may be negative and that there are even cases where it is equal to $-\\infty$ (e.g., when $\\textnormal{dist}(A,\\Omega)>0)$; see \\cite{LapRaZu}.}\n\n\\begin{remark}\n\\label{rmk:StandardFractalDrum}\nFor the standard notion of fractal drum, which corresponds to the choice $A=\\partial\\Omega$ with $\\Omega$ bounded, we refer, e.g., to \\cite{Lap1,Lap2} along with \\cite[\\S 12.5]{LapvF6} and the relevant references therein. \n\\end{remark}\n\n\\begin{remark}\n\\label{rmk:RelativeFractalDrums}\nIt is easy to see that the notion of relative fractal drum $(A,\\Omega)$ is a natural extension of the notion of fractal string ${\\mathcal{L}}=\\{\\ell_j\\}$. Indeed, for a given (standard) fractal string ${\\mathcal{L}}=\\{\\ell_j\\}$, it suffices to define $A=\\{a_j\\}$, where $a_j:=\\sum_{k\\ge j}\\ell_k$ and $\\Omega=\\cup_{k\\ge1} (a_{k+1},a_k)$. We caution the reader that the notion of generalized fractal string already exists but does not coincide with the notion of relative fractal drum. Specifically, in \\cite[Ch.~4]{LapvF6}, a generalized fractal string is defined to be a locally positive or locally complex measure on $(0,\\infty)$ supported on a subset of $(x_0,\\infty)$, for some positive real number $x_0$.\n\\end{remark}\n\n\n\\section{Summary of results and open problems}\n\\label{sec:SummaryOfResultsAndOpenProblems}\n\nFor a bounded infinite set $A$, recall that $\\overline{\\dim}_{B}A$ denotes the upper box-counting dimension of $A$ given by Equation \\eqref{eqn:UpperBoxDimension}, $\\overline{\\dim}_{M}A$ denotes the upper Minkowski dimension of $A$ given by Equation \\eqref{eqn:UpperMinkowskiDimension}, $D_B$ denotes the abscissa of convergence of the box-counting zeta function $\\zeta_B$ of $A$ given in Definition \\ref{def:BoxCountingZetaFunction}, $\\rho_\\mathcal{L}$ denotes the order of the geometric counting function $N_\\mathcal{L}$ given by Equation \\eqref{eqn:OrderOfCountingFunction} where $\\mathcal{L}=\\mathcal{L}_B$, $D_N$ denotes the value corresponding to the (asymptotic) growth rate of $N_\\mathcal{L}$ given by Equation \\eqref{eqn:CountingFunctionD}, and $D_d$ is the abscissa of convergence of the distance zeta function $\\zeta_d$ given in Definition \\ref{def:DistanceZetaFunction}. Furthermore, let $D_t$ denote the abscissa of convergence of the tube zeta defined in Equation \\eqref{tube_zeta}.\n\nThe following theorem summarizes our main result (as stated in Theorem \\ref{thm:Summary} of the introduction), which pertains to the determination of the box-counting dimension of a bounded infinite set. (Recall that the equalities $\\overline{\\dim}_{B}A=D_d=D_t$ are established in \\cite{LapRaZu}; see Theorem \\ref{thm:DistanceZetaFunctionUpperBoxDimension} and the comment following Theorem \\ref{thm:DistanceAndTubeZetaFunctions} above.)\n\n\\begin{theorem}\n\\label{thm:MainResult}\nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$ and let $\\mathcal{L}=\\mathcal{L}_B$ be the corresponding box-counting fractal string. Then the following equalities hold\\emph{:}\n\\begin{align*}\n\\overline{\\dim}_{B}A &= \\overline{\\dim}_{M}A = D_B = \\rho_\\mathcal{L} = D_N = D_d=D_t.\n\\end{align*} \n\\end{theorem}\n\n\\begin{proof}\nThe classic equality $\\overline{\\dim}_{B}A = \\overline{\\dim}_{M}A$ is established in \\cite{Falc}. The equality $\\overline{\\dim}_{M}A=D_B=D_N$ follows from Theorem \\ref{thm:JohnnyDarko}. The equality $\\rho_\\mathcal{L}=D_N$ follows from Remark \\ref{rmk:DimensionEqualsOrder}. Finally, as was recalled just above, the equalities $\\overline{\\dim}_{M}A=D_d$ and $\\overline{\\dim}_{M}A=D_t$ are established in \\cite{LapRaZu}; see Theorem \\ref{thm:DistanceZetaFunctionUpperBoxDimension} and the comment following Theorem \\ref{thm:DistanceAndTubeZetaFunctions}. These last two equalities are valid whether or not $A$ is infinite.\n\\end{proof}\n\nRecall that, as stated in Definition \\ref{def:BoxCountingFunction}, $N_B(A,x)$ denotes the maximum number of disjoint balls of radius $x^{-1}$ centered in $A$. In this setting and for $\\varepsilon>0$ we have\n\\begin{align}\n\\label{eqn:LowerBoundForVolume}\nB_m \\varepsilon^m N_B(A,\\varepsilon^{-1}) \\leq |A_\\varepsilon|,\n\\end{align}\nwhere $A_\\varepsilon$ is the $\\varepsilon$-neighborhood of $A$, $|A_\\varepsilon|=|A_\\varepsilon|_m$ $B_m$ is the $m$-dimensional volume of a ball in $\\mathbb{R}^m$ with unit radius, and $0<\\varepsilon<x_1^{-1}$, where $x_1^{-1}$ is given by Proposition \\ref{prop:BoxCountingJumps}. \n\nMotivated by Equation \\eqref{eqn:LowerBoundForVolume} and Theorem \\ref{thm:CountingFunctionAndComplexDimensions}, we propose the following open problem (which is stated rather roughly here).\n\n\\begin{openproblem}\n\\label{op:BoxCountingAndCountingFormulas}\nLet $A$ be a bounded infinite subset of $\\mathbb{R}^m$ with box-counting fractal string $\\mathcal{L}_B$. Assume suitable growth conditions on $\\zeta_B$ \\emph{(}such as the languidity of $\\zeta_B$ on an appropriate window, see \\emph{\\cite[Chs.~5 \\& 8]{LapvF6})} and assume for simplicity that all of the complex dimensions are simple \\emph{(}i.e., are simple poles of $\\zeta_B$\\emph{)}. Then, as $\\varepsilon\\to 0^+$, compare the quantities\n\\begin{align}\n\\label{eqn:VolumeFormulaOpenProblem}\n|A_\\varepsilon|, \\quad \\varepsilon^m N_{\\mathcal{L}_B}(\\varepsilon^{-1}), \\quad \\textnormal{and } \\, \\varepsilon^m \\left(\\sum_{\\omega \\in \\calD_B}\n\\frac{\\varepsilon^{-\\omega}}{\\omega}\\textnormal{res}(\\zeta_B(s);\\omega) +  R(\\varepsilon^{-1})\\right), \n\\end{align}\nwhere $R(\\varepsilon^{-1})$ is an error term of small order. \n\\end{openproblem}\n\nIf one were to provide a more precise version of the above open problem and solve it, one might consider pursuing a generalization of Theorem \\ref{thm:CriterionForMinkowskiMeasurability} in the spirit of the theory of complex dimensions of fractal strings, as described in \\cite{LapvF6}, and of its higher-dimensional counterpart in \\cite{LapPe3,LapPe2,LapPeWin}. Naturally, the clarified version of this open problem would consist of replacing the implicit `approximate equalities' in Equation \\eqref{eqn:VolumeFormulaOpenProblem} with true equalities, modulo suitable modifications  and under appropriate hypotheses.\n\nAnalogously (but possibly more accurately), in light of the results from \\cite{LapRaZu} discussed in Section \\ref{sec:DistanceAndTubeZetaFunctions}, as well as from the results about fractal tube formulas obtained in \\cite[Ch.~~8]{LapvF6} for fractal strings and in \\cite{LapPe3,LapPe2} and especially \\cite{LapPeWin} in the higher-dimensional case (for fractal sprays and self-similar tilings),\\footnote{A survey of the results of \\cite{LapPe3,LapPe2,LapPeWin} can be found in \\cite[\\S 13.1]{LapvF6}.} we propose the following open problem. (A similar problem can be posed for the tube zeta function $\\tilde\\zeta_A$ discussed in Section \\ref{sec:TubeZetaFunction}.)\n\n\\begin{openproblem}\n\\label{op:VolumeTubularNeighborhood}\nLet $A$ be a bounded subset of $\\mathbb{R}^m$ with distance zeta function $\\zeta_d$. Under suitable growth assumptions on $\\zeta_d$ \\emph{(}such as the languidity of $\\zeta_B$ on an appropriate window, see \\emph{\\cite[Chs.~5 \\& 8]{LapvF6})}, and assuming for simplicity that all of the corresponding complex dimensions are simple, calculate the volume of the tubular neighborhood of $A$ in terms of the complex dimensions of $A$ \\emph{(}defined here as the poles of the meromorphic continuation of $\\zeta_d$ union the `integer dimensions' \\{0,1,\\ldots,m\\}\\emph{)} and the associated residues. \n\nMoreover, even without assuming that the complex dimensions are simple, express the resulting fractal tube formula as a sum of residues of an appropriately defined `tubular zeta function' \\emph{(}in the sense of \\emph{\\cite{LapPe3,LapPe2,LapPeWin,LapPeWi2})}.\n\\end{openproblem}\n\n\\section*{Acknowledgments}\n\nThe first author (M.~L.~Lapidus) would like to thank the Institut des Hautes Etudes Scientifiques (IHES) in Bures-sur-Yvette, France, for its hospitality while he was a visiting professor in the spring of 2012 and this paper was completed.\n\nThe authors would like to thank our anonymous referees. They provided helpful comments and suggestions in their very thorough reviews of a preliminary version of this paper.\n\nIn closing, the authors would like to thank the Department of Mathematics at the University of Messina and the organizers of the Permanent International Session of Research Seminars (PISRS), especially Dr.\\ David Carfi, for their organization of and invitations to speak in the First International Meeting of PISRS, Conference 2011: Analysis, Fractal Geometry, Dynamical Systems, and Economics. The authors' participation in this conference led directly to collaboration on the development of the new results presented in the paper, namely the results pertaining to box-counting fractal strings and box-counting zeta functions. \n\n\n\\bibliographystyle{amsalpha}\n\n", "meta": {"timestamp": "2013-02-04T02:00:32", "yymm": "1207", "arxiv_id": "1207.6681", "language": "en", "url": "https://arxiv.org/abs/1207.6681", "abstract": "We discuss a number of techniques for determining the Minkowski dimension of bounded subsets of some Euclidean space of any dimension, including: the box-counting dimension and equivalent definitions based on various box-counting functions; the similarity dimension via the Moran equation (at least in the case of self-similar sets); the order of the (box-)counting function; the classic result on compact subsets of the real line due to Besicovitch and Taylor, as adapted to the theory of fractal strings; and the abscissae of convergence of new classes of zeta functions. Specifically, we define box-counting zeta functions of infinite bounded subsets of Euclidean space and discuss results pertaining to distance and tube zeta functions. Appealing to an analysis of these zeta functions allows for the development of theories of complex dimensions for bounded sets in Euclidean space, extending techniques and results regarding (ordinary) fractal strings obtained by the first author and van Frankenhuijsen.", "subjects": "Mathematical Physics (math-ph); Number Theory (math.NT)", "title": "Box-counting fractal strings, zeta functions, and equivalent forms of Minkowski dimension", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9861513889704252, "lm_q2_score": 0.8596637451167997, "lm_q1q2_score": 0.8477585962944496}}
{"url": "https://arxiv.org/abs/2301.08149", "title": "Zeros of fractional derivatives of polynomials", "abstract": "We investigate the behavior of fractional derivatives of polynomials. In particular, we consider the locations and the asymptotic behaviour of their zeros and give bounds for their Mahler measure.", "text": "\\section{Introduction}\n\nQuestions concerning finding exact or approximate values of the zeros of polynomial functions $p(x) = c_n x^n + c_{n-1}x^{n-1} + \\cdots + c_1 x + c_0$ are classical, and (for the case of real coefficients $c_0, c_1, \\cdots c_n$) properties of the distribution of these zeros have been studied since at least 1637, when Descartes established his fundamental Rule of Signs (in {\\em La G{\\' e}om{\\' e}trie} \\cite{descartes1637}). This important result was refined to finite intervals by Budan in 1807 and by Fourier in 1820 (see \\cite{bf1} or \\cite{bf2}). By then, thanks to the work of Euler, Gauss and Argand (see \\cite{fine1997} for more details), the Fundamental Theorem of Algebra had been established, guaranteeing that a polynomial of degree $n$ has exactly $n$ complex zeros (counted with multiplicity). Not much later,\nin 1829, Cauchy \\cite{cauchy1829} was able to prove that all zeros of a monic polynomial $p(x)$, with complex coefficients, must lie inside the disk $|z| < 1 + \\max_{0 \\leq k \\leq n-1} |c_k|$, \nthe bounds that were eventually generalized by Landau \\cite{landau1907}, Fej{\\' e}r \\cite{fejer1908} and others.  \n\nIn another direction,\none could ask about the relation between locations of the zeros of a polynomial $p(x)$ and the zeros of its derivative $p^{\\prime}(x)$, as Rolle has done in his {\\em  Trait{\\' e} d’alg{\\`e}bre} of 1690 (see \\cite{shain1937}); the well-known theorem bearing his name -- that states that between any two zeros of a real polynomial there lies at least one zero of the derivative -- was proved rigorously by Cauchy \\cite{cauchy1823} in 1823. In the complex plane, the situation becomes even more interesting. As Gauss noted in 1836, all zeros of $p^{\\prime}(x)$ lie in the convex hull of the zeros of $p(x)$.\nThe first proof of this proposition was published by Lucas \\cite{lucas1874} in 1874; it is now known as the Gauss-Lucas Theorem\n(also see \\cite{marden1966}). At the beginning of the 20th century it was refined by B{\\^o}cher \\cite{bocher1904}, Jensen \\cite{jensen1912} and Walsh \\cite{walsh1920}, and in more recent times, several related extensions and generalizations of it have been considered by Dimitrov \\cite{dimitrov1998}, Brown \\& Xiang \\cite{brown1999}, Sendov \\cite{sendov2021}, Tao \\cite{tao2022} and others.\n\n \n\nThe main aim of our work is to investigate connections between these two central themes. We will try to show that their key ideas can be combined in a very natural way, but to quite surprising effects, if one considers the fractional \nderivatives $p^{(\\alpha)} (x)$, where $\\alpha \\in \\mathbb{R}$ is a variable $0 \\leq \\alpha \\leq n = \\deg p(x)$. \nOur main goal in this paper will be to answer one of the most intriguing questions that arises as soon as one begins to study these topics: since obviously $\\deg p^{\\prime}(x) =  \\deg p(x) - 1$, and the Fundamental Theorem asserts than the same reduction must occur for the total number of zeros, \nwhat happens to the zeros of fractional derivatives, as the real $\\alpha$ increases continuously from $0$ to $n$? How \ndo the zeros of polynomials vanish, and why? As it turns out, these questions have remarkably simple and elegant answers. Namely, for a polynomial $p(x)$ of degree $n$, each of its $n$ zeros will belong to a path of unique length that connects it to the origin, where the ``length'' of the path can be measured by the number of zeros of its derivatives it contains; in other words, for each $0 \\leq k \\leq n-1$ there will be a unique path (originating at one of the zeros of $p(x)$) that will contain exactly $k$ zeros of its higher derivatives. (Figure \\ref{fig two} shows this general property for a generic cubic polynomial).  \n\n \n\n\\begin{figure}[ht]\n\\includegraphics[width=0.45\\textwidth]{paper-deg-3-center-0.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-3-center+1-I.pdf}\n\\caption{Paths \\(z(\\alpha)\\) of zeros of the Riemann-Liouville fractional derivatives \n\\(\\RL{0}{\\alpha}p\\) and\n\\(\\RL{1-i}{\\alpha}p\\)\nof the polynomial \\(p(x)=x^3 + (1+i)x^2 + x - i\\).\nWe end the paths when \\(z(\\alpha)\\) reaches the ``origin'' (\\(a=0\\) in the first case, and \\(a=1-i\\) in the second).\n}\\label{fig two}\n\\end{figure}\n\nAnother goal of this paper will be to try to understand some of the particulars of the paths the fractional zeros take, their dynamical properties. In order to state our results concerning this general flow of polynomial zeros more precisely, first we will need to recall some basic definitions and properties \nof fractional derivatives. There exist a multitude of different definitions of fractional derivatives, each with its own particular advantages and disadvantages. Unlike in \nour study of the Riemann zeta function and the Stieltjes constants where the Gr\\\"unwald-Letnikov fractional derivative\n\\cite{g:1867,l:1869-1} worked \nreally well (see \\cite{fps} and \\cite{fps2}), in the case of polynomials their divergence forces us instead to employ \nthe  Riemann-Liouville fractional derivative \\cite{riefrac}, which also can be thought of as a truncated version of the Gr\\\"unwald-Letnikov fractional derivative \\cite{ortigueira2011fractional}. \n\n\nThe rest of the paper is structured as follows. Our Section \\ref{sec derv} contains a short introduction to the Riemann-Liouville differintegral  focusing on results most applicable to fractional derivatives of polynomials. \nIn Section \\ref{sec low} this theory is applied to the two simplest cases: polynomials of degree one and two. These are the two cases where, thanks to the manageable classical formulas for the zeros, all the main questions can be conclusively answered. With the cubic polynomials things become somewhat murky, but general convergence trends can still be established. In Section \\ref{sec flow} we do just that: we examine how the zeros of integral derivatives are connected to the zeros of fractional derivatives in the most general setting, and we look at the paths of zeros and investigate their convergence and the overall flow. In Section \\ref{sec bounds} we consider the behaviour of the zeros on a larger scale and we prove bounds for\nthe Mahler measure of the fractional derivatives, which are then, in Section \\ref{caputo}, also established for the Caputo fractional derivative. Finally, in Section \\ref{conclusion}, we discuss some intriguing open problems and unsolved questions. \n\n\\section{Riemann-Liouville Fractional \nDerivatives}\\label{sec derv}\nIn full generality, the \\(\\alpha\\)-th Riemann-Liouville fractional derivative is defined as follows, see \\cite{oldham-spanier}, for example: \n\n\\begin{definition}[Riemannn-Liouville fractional derivative]\\label{def rl}\nLet \\(f\\) be analytic on a convex open set \\(C\\) let and \\(a\\in C\\). \nFor \\(\\alpha>0\\) the Riemann-Liouville integral is\n\\[\n\\RLI{a}{\\alpha} f(t)=\\frac{1}{\\Gamma(\\alpha)}\n\\int_a^t f(t)(t-\\tau)^{\\alpha-1} \\,d\\tau.\n\\]\nSet \\(\\RLI{a}{0}f(t)=f(t)\\) and for \\(\\alpha>0\\) define the Riemann-Liouville fractional derivative as\n\\[\n\\RL{a}{\\alpha} f(t) = \\frac{d^m}{dt^m} \\RLI{a}{m-\\alpha} f(t),\n\\]\nwhere $m=\\lceil \\alpha \\rceil$.  For \\(\\alpha<0\\) set\n\\(\\RL{a}{\\alpha}f(t)=\\RLI{a}{-\\alpha} f(t)\\).\n\\end{definition}\nIn what follows, we will only consider the special case of polynomials, composed of the simple power functions $p(x)=(x-a)^{\\beta}$, where $\\beta \\in \\mathbb{R}$, $a\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}$. For these, the $\\alpha$-th Riemann-Liouville fractional derivative can be computed \nusing the Power Rule: \n\\begin{align}\\label{eq poly derv}\n\\RL{a}{\\alpha}(x-a)^{\\beta}=\\begin{cases}\n0 & \\text{if }\\alpha-\\beta-1 \\in \\N\\\\\n\\dfrac{\\Gamma(\\beta+1)}{\\Gamma(\\beta-\\alpha+1)}(x-a)^{\\beta-\\alpha} & \\text{otherwise}\n\\end{cases}\n\\end{align}\nwhere, for $z \\in \\mathbb{C}$, the gamma function $\\Gamma(z)$ is defined as $\\Gamma(z)=\\int_0^\\infty t^{z-1}e^{-t}dt$; it satisfies $\\Gamma(1+n)=n\\Gamma(n)$, which implies that $\\Gamma(n) = (n-1)!$ for $n \\in \\mathbb{N}$. The function $\\Gamma(z)$ has no zeros in the complex \nplane $\\mathbb{C}$, and has poles at all the negative integers (see \\cite{artin1964}).\n\n\\begin{remark}\nThe Riemann-Liouville fractional derivative of a monomial \\(f(x)=x^n\\) is multivalued.  When changing the branch of the complex logarithm in the computation of the fractional derivative all coefficients of the derivative are changed by the same factor.\nSo choosing a different branch of the complex logarithm does not change the zeros of the derivative, which means that we\ncan fix the branch in our consideration of zeros of derivatives of polynomials.\nWe use the principal branch of the complex logarithm.\n\\end{remark}\n\nIt should be noted that the constant $a$ that centers the expansion (\\ref{eq poly derv}) plays a key role in all our computations below, as the ``origin,'' or the limit of convergence, of the flow of zeros of derivatives (Figure \\ref{fig two} illustrates its role). Also noteworthy is the fact that the Riemann-Liouville fractional derivative satisfies all properties expected of a regular derivative, with the exception of the composition rule. The following example shows why it fails: \n\n\\begin{example} \nFrom (\\ref{eq poly derv}),  the \\(1.5\\)-th derivative of \\(p(x)=1\\) is\n\\(\n\\textstyle\\RL{0}{1.5}p(x)=\\frac{2}{ \\csqrt{\\pi}}x^{-1.5}\n\\)\nand \n\\(\\textstyle\\RL{0}{1}\\left(\\RL{0}{0.5}p(x)\\right) =\\frac{2}{ \\csqrt{\\pi}}x^{-0.5}.\n\\)\nHowever, \\(\\RL{0}{0.5}\\left(\\RL{0}{1}p(x)\\right)=\\RL{0}{0.5}0=0.\n\\)\n\\end{example}\n\nWhen \\(\\beta\\in\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}\\setminus\\Z\\) we still have\n\\(\\RL{a}{\\alpha}\\left(\\RL{a}{\\beta}p(x)\\right)=\\RL{a}{\\alpha+\\beta}p(x)\\).\n\n\n\n\\begin{remark} It is possible to go beyond the standard values of $0 \\leq \\alpha$, and consider what happens for $\\alpha < 0$. Here, there will be \\(n+1\\) complex roots, because the first term in Equation \\ref{eq lem derv} below has the root \\(0\\). Just like in the standard case, the extended curves \\(z(\\alpha)\\) of zeros of the differintegral \\(\\RL{a}{\\alpha}p\\) will be continuous for \\(\\alpha<0\\) unless\n\\(\\RL{a}{\\alpha}p\\) has a double root; however, they will not be smooth at integral \\(\\alpha > 0\\).  More on this will be said in Section \\ref{sec flow} below.\n\\end{remark}\n\n\n\n\nIn what follows, we consider the zeros of the the fractional derivatives of polynomials \\(p\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) of degree \\(n\\), and we investigate the implicit functions \\(z:(0,n)\\to \\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}\\) given by\n\\[\n\\RL{a}{\\alpha}f(z(\\alpha))=0.\n\\]\nIf \\(\\left(\\RL{a}{\\alpha}p(x)\\right)'\\ne 0\\), for \\(\\alpha\\in(0,n)\\), then \n\\(z(\\alpha)\\) is differentiable on \\((0,n)\\).\nWe denote the roots of the polynomial \\(p(x)\\) by \\(z_1, z_2, \\dots,z_n\\)\nand for \\(1\\le k\\le n\\) we define the the implicit function \\(z_k:[0,n)\\to\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}\\)  by \\(z_k(0)=z_k\\)\nand \\(\\RL{a}{\\alpha}p(z(\\alpha))=0\\).\n\n\n \n\nThe following representation of the fractional derivatives of a general monic polynomial will be most useful.\n\n\\begin{lemma}\\label{lem derv}\nLet \\(p(x)=(x-a)^n+\\sum_{j=0}^{n-1} c_j (x-a)^{j}\\in\\mathbb{C}[x]\\) and \\(\\alpha\\not\\in\\N\\).  Then\n\\begin{equation}\\label{eq lem derv}\n\\RL{a}{\\alpha}p(x)=\\frac{n!}{\\Gamma(n+1-\\alpha)} (x-a)^{-\\alpha}\n\\left[(x-a)^{n}+\n\\sum_{j=0}^{n-1} \n\\left(\n\\prod_{k=j+1}^n\\!(k-\\alpha)\\right)\\cdot\n\\frac{j!}{n!} c_j (x-a)^{j}\n\\right].\n\\end{equation}\n\\end{lemma}\n\n\n\\begin{proof}\nApplying the Power Rule (\\ref{eq poly derv}), we obtain \\begin{align*}\n\\RL{a}{\\alpha}p(x)  &=\n\\frac{\\Gamma(n+1)}{\\Gamma(n+1-\\alpha)} (x-a)^{n-\\alpha}+\n\\sum_{j=0}^{n-1} \\frac{\\Gamma(j+1)}{\\Gamma(j+1-\\alpha)} c_j (x-a)^{j-\\alpha}\\\\\n&=\\frac{\\Gamma(n+1)}{\\Gamma(n+1-\\alpha)} (x-a)^{-\\alpha}\n\\left[(x-a)^{n}+\n\\sum_{j=0}^{n-1} \n\\frac{\\Gamma(n+1-\\alpha)}{\\Gamma(j+1-\\alpha)}\n\\frac{\\Gamma(j+1)}{\\Gamma(n+1)} c_j (x-a)^{j}\n\\right]\n\\\\\n&=\\frac{n!}{\\Gamma(n+1-\\alpha)} (x-a)^{-\\alpha}\n\\left[(x-a)^{n}+\n\\sum_{j=0}^{n-1} \n\\left(\n\\prod_{k=j+1}^n\\!(k-\\alpha)\\right)\\cdot\n\\frac{j!}{n!} c_j (x-a)^{j}\n\\right],\n\\end{align*}\nas wanted. \\qedhere\n\\end{proof}\n\n\\begin{remark} The representation of the fractional derivatives of polynomials given in the above  Lemma \\ref{lem derv} has the property that their roots only depend on the factors in square brackets, which in turn implies the useful fact that the branch cut of the complex logarithm does not affect the paths \\(z(\\alpha)\\) of\nzeros of these fractional derivatives.\n\nWe also see that for \\(p(x)=(x-a)^n\\) we have:\n\\begin{equation}\\label{eq at a}\n\\RL{a}{\\alpha}p(a)=\n\\begin{cases}\n0 &\\mbox{if }\\alpha<n\\\\\n\\Gamma(n+1) & \\mbox{if }\\alpha=n\\\\\n\\mbox{undefined} &\\mbox{if }\\alpha>n\n\\end{cases}\n\\end{equation}\n\\end{remark}\n\n\n\\begin{remark} A few words should be said about our plots of the implicit functions \\(z:\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}\\to\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}\\)\nwith \\(z(\\alpha)\\) given by \\(\\RL{a}{\\alpha}p(z(\\alpha))=0\\).  \nThe dots labelled `\\(\\bullet\\)\\textsf{k}' represent zeros of the \\(k\\)th Riemann-Liouville differintegral (thus `\\(\\bullet \\)\\textsf{0}' \nrepresent the zeros of the polynomial \\(p(x)\\) itself), while circles\n`\\(\\circ\\)\\textsf{k}' represent points that are limits of \\(z(\\alpha)\\) as \\({\\alpha\\to k}\\) but are not zeros of \\(\\RL{a}{k}p\\).  These occur for integral \\(k\\) with \\(k\\ge\\deg p(x)\\), where \\(\\RL{a}{k}p(x)\\) is constant, for example at \\(x=a\\), see Equation (\\ref{eq at a}).\nMoreover, when a point is either a zero or a limit point of zeros of both the \\(j\\)th and the \\(k\\)th differintegrals, then it is represented by `\\(\\bullet\\) \\textsf{j \\& k}' or `\\(\\circ\\) \\textsf{j \\& k}' respectively.  \nIn Figures \\ref{fig two}, \\ref{fig double}, and \\ref{fig longest} we let all paths of zeros of \\(\\RL{a}{\\alpha}p(x)\\) end when they reach the origin \\(a\\).  In Figures \\ref{fig one},  \\ref{fig double}, \\ref{fig p1},\n\\ref{fig quintic paths}, and \\ref{fig quintic abs}\nwe continue the paths past \\(a\\). \nIn Figures \\ref{fig one}, \\ref{fig double}, and \\ref{fig p1} we display the path of zeros \\(z(\\alpha)\\) for \\(\\alpha<0\\) and \\(\\alpha>n\\) in lighter colors than for \\(0\\le \\alpha < n \\) where \\(n\\) is the degree of the polynomial.  \nThe point \\(a\\) is only labeled with values for \\(\\alpha\\) for \\(\\alpha\\ge 0\\).\n\\end{remark}\n\n\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-1.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-2.pdf}\n\\caption{Path $z(\\alpha)$ of zeros of differintegrals of the polynomials \\(p(x)=x-2-3i\\) and $p(x)=(x-2-3i)(x+2+i)$.}\n\\label{fig one}\n\\end{figure}\n\n\n\\section{Low Degrees}\\label{sec low}\nFormulas for finding zeros of polynomials of low degrees have been known for centuries. Applying the Riemann-Liouville derivative to these low-degree cases\nhas proved to be a simple but informative exercise. In this section we summarize some of these results, stating the most useful ones as lemmas. They are \nexamples of a \ndynamic that shares certain key characteristic with most high-degree cases, but some aspects of which are often unique. For example, in the linear case, \nthe path the zeros take is also linear, while already in the quadratic case one observes a considerably more complex behavior.\n\nLet us start with the linear polynomials. Here the situation is simple. The paths of zeros will always be linear, too, and they can be completely described. We have: \n\n\\begin{lemma}[Linear Polynomials]\\label{lem linear}\nLet \\(p(x)=(x-a)+c_0\\). As $\\alpha$ increases from $0$ to $1$, the path of zeros of \\(\\RL{a}{\\alpha}p(x)\\) is given by \\(z(\\alpha)=(\\alpha-1)c_0+a\\).\n\\begin{proof}\nFrom (\\ref{eq poly derv}), with $\\beta = 1$, we get: \\(\\RL{a}{\\alpha}p(x)=(x-a)^{-\\alpha}\\left((x-a)+(1-\\alpha)c_0\\right)\\).\n\\end{proof}\n\\end{lemma}\n\n\\begin{remark}\n    In addition to considering the $\\alpha$-th derivatives in the usual range $0 \\leq \\alpha \\leq n$, one could also look at what happens when $\\alpha < 0$ and when $\\alpha>n$.   \n    In the linear case this is, again, simple.   From Lemma \\ref{lem linear} we can deduce that, as with $\\alpha<0$, the line of roots of the derivatives continues. Similarly, for $\\alpha>1$, $\\alpha\\not\\in\\Z$ lie on the same line, see Figure \\ref{fig one} below. \n\\end{remark}\n\n\n\n\n\n\n\n\n\n\n\nLet us now consider the quadratic case. This case is considerably more interesting, since there are\nnow two paths of zeros of the fractional derivatives, and they exhibit a much more complex and intricate behavior.\nWe first notice that the path of the zero closest to $a$ will directly connect with $a$, while the path of the farther zero will in the process of reaching $a$ pass through the zero of the first derivative. \n\n\\begin{lemma}\\label{lem quad root order}\nLet $p(x)=x^2+c_1(x-a)+c_0$ with discriminant \\(d=\\) and roots \\(s_1\\) and \\(s_2\\).\n\\begin{enumerate}\n\\item If \\(d\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}\\setminus\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}^{<0}\\), denote by \\( \\csqrt{d}\\) the \ncomplex number $r$ with \\(r^2=d\\) and \\(\\Re(r)> 0\\).  Then for the \nroots \\(s_1=\\frac{-b+ \\csqrt{b^2-4c}}{2}\\) and \\(s_2=\\frac{-b- \\csqrt{b^2-4c}}{2}\\)\nof \\(g(x)=x^2+bx+c\\in\\mathbb{C}[x]\\) we have \\(|s_1|\\le |s_2|\\).\n\\item If \\(d\\in\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}^{<0}\\), then \\(|s_1|=|s_2|\\).\n\\end{enumerate}\n\\end{lemma}\n\n\\begin{proof}\n\\begin{enumerate}\n\\item Write \\(d=1-\\frac{4c}{b^2}\\).\nThen\n\\(|s_1|=\\left|\\frac{b}{2}\\right|\\cdot \\left|-1+ \\csqrt{d}\\right|\\)\nand \\(|s_2|=\\left|\\frac{b}{2}\\right|\\cdot \\left|-1- \\csqrt{d}\\right|\\).\nNow\n\\begin{align*}\n\\left|-1+ \\csqrt{d}\\right|^2 & = \\left(-1+ \\csqrt{d}\\right)\\left(-1+\\overline{ \\csqrt{d}}\\right)=1- \\csqrt{d}-\\overline{ \\csqrt{d}}+|d| \\text{ and }\\\\\n\\left|-1- \\csqrt{d}\\right|^2 &= \\left(-1- \\csqrt{d}\\right)\\left(-1-\\overline{ \\csqrt{d}}\\right)=1+ \\csqrt{d}+\\overline{ \\csqrt{d}}+|d|\n\\end{align*}\nBecause \\(\\Re\\left( \\csqrt{d}\\right)> 0 \\) we have\n\\(|s_1|\\le |s_2|\\). \n\\item Now assume $\\Re(r)=0$, meaning $r$ is completely imaginary, i.e. $r=it$. Then using the information above we have, \n\\begin{align*}\n    \\left|-1+ \\csqrt{d}\\right|^2 & =1- \\csqrt{d}-\\overline{ \\csqrt{d}}+|d| = 1-r-\\overline{r}+|r^2| =1-it-\\overline{it}+\\left|(it)^2\\right| \\\\\n    & =1-t^2 =1+it-it-t^2 =1+it+\\overline{it}-t^2 =\\left|-1- \\csqrt{d}\\right|^2.\n\\end{align*}\nTherefore, we have \\(|s_1|= |s_2|\\).\\qedhere\n\\end{enumerate}\n\\end{proof}\n\n\\begin{remark}\n    As of yet, there is no known reliable ordering of the zeros for any of the higher degree polynomials. In fact, there exist examples of cubic polynomials for which the standard Euclidean distance (which works so well for the linear and the quadratic cases) can be shown to fail: see Figure \\ref{fig longest}.  \n\\end{remark}\n\n\n \n\n In addition to the natural ordering on the quadratic roots, another question that seems to be of interest is the one that concerns the trends of descent of their paths, especially since it had such a nice answer in the linear case. As it turns out, the asymptotes of the two quadratic paths exist, and the quadratic formula alone is enough to help us find them. \n \n  \n \n\n\\begin{theorem}\\label{prop roots quad}\nFor the quadratic polynomial $p(x)=(x-a)^2+c_1(x-a)+c_0$, the paths of zeros of the fractional derivatives \\(\\RL{a}{\\alpha} p(x)\\) are given as\n\\begin{equation}\\label{eq roots quad}\nz_{1,2}(\\alpha)=a+\\frac{-(2-\\alpha) c_1\\pm \\csqrt{(2-\\alpha)^2\\cdot (c_1)^2 - 8(2-\\alpha)(1-\\alpha) \\cdot c_0}}{4},\n\\end{equation}\nwith \\(|z_1(\\alpha)|\\ge|z_2(\\alpha)|\\), for {\\(\\alpha\\in[0,1]\\)}, and \\(\\lim_{\\alpha\\to 1}z_1(\\alpha)=a\\)\nand \n\\(\\lim_{\\alpha\\to 2}z_{1,2}(\\alpha)=a\\).\n\\begin{proof}\nWith the help of (\\ref{eq poly derv}), the fractional derivatives of \\(p(x)\\) can be written as \n\\[\n\\RL{a}{\\alpha}p(x)=\\frac{2}{\\Gamma(3-\\alpha)}x^{-\\alpha}\n\\left(\n(x-a)^{2} +\\frac{(2-\\alpha)\\cdot c_1}{2} (x-a) +\\frac{(2-\\alpha)(1-\\alpha)\\cdot c_0}{2}\\right).\n\\]\nNow, set $y=x-a$. Then the roots of $y^{2} +\\frac{(2-\\alpha)\\cdot c_1}{2} y +\\frac{(2-\\alpha)(1-\\alpha)\\cdot c_0}{2}$ . \nare \n\\[\nz_{1,2}(\\alpha)=a+\\frac{-(2-\\alpha) c_1\\pm \\csqrt{(2-\\alpha)^2\\cdot (c_1)^2 - 8(2-\\alpha)(1-\\alpha) \\cdot c_0}}{4}\n\\]\nThe ordering of the roots \\(|z_1(\\alpha)|\\ge|z_2(\\alpha)|\\), for {\\(\\alpha\\in[0,1]\\)},  follows with Lemma \\ref{lem quad root order}. Furthermore, we have \n\\[\n\\lim_{\\alpha\\to 1} z_{1,2}(\\alpha)=a+\\frac{- c_1\\pm \\csqrt{ (c_1)^2}}{4}=a+\\frac{-c_1\\pm c_1}{4}.\n\\]\nThus $\\lim_{\\alpha\\to 1} z_1(\\alpha)=a$ and $\\lim_{\\alpha\\to 1} z_2(\\alpha)= a+\\frac{c_1}{2}$.\nSimilarly\n\\[    \n\\lim_{\\alpha\\to 2}\\left[a+\\frac{-(2-\\alpha) c_1\\pm \\csqrt{(2-\\alpha)^2\\cdot (c_1)^2 - 8(2-\\alpha)(1-\\alpha) \\cdot c_0}}{4}\\right]=a.\\qedhere\n\\]\n\\end{proof}\n\\end{theorem}\n\n \n\n\\begin{corollary}\nFor \\(\\alpha \\to \\pm\\infty\\), the asymptotes of the quadratic paths are\n\\[\\textstyle z_{1,2}(\\alpha)\\approx\\frac{(2-\\alpha) c_1}{4}\\left[-1\\pm \\csqrt{1-\\frac{8c_0}{c_1}}\\right]. \\]\n\\end{corollary}\n\n\\begin{proof}\nBy (\\ref{eq roots quad}) \nwe have,   \n\\begin{align*}\n        z_{1,2}(\\alpha)&=\\frac{-(2-\\alpha) c_1\\pm(2-\\alpha) c_1 \\csqrt{1 - 8\\frac{(1-\\alpha)}{(2-\\alpha)} \\cdot \\frac{c_0}{c_1}}}{4}\\\\\n        &=\\frac{(2-\\alpha) c_1}{4}\\left[-1\\pm \\csqrt{1-\\frac{8c_0(1-\\alpha)}{c_1^2(2-\\alpha)}}\\right],\\\\\n\\end{align*}\nand since $\\frac{1-\\alpha}{2-\\alpha}\\to 1$, as $\\alpha\\to \\pm\\infty$, this yields linear asymptotes for $z_1(\\alpha)$ and $z_2(\\alpha)$.\n\\end{proof}\n\n\n\nAn noteworthy special case occurs when the polynomial has a double root. Then the paths display an interesting symmetry, see Figure \\ref{fig double}.\nIn fact, it is easy to see that specializing our Proposition \\ref{prop roots quad} to the case of a double zero of the polynomial itself yields:\n\n\\begin{corollary}\nIf $p(x)=(x-z_0)^2$ then the zeros of \\(\\RL{0}{\\alpha}p\\) are\n\\[z_{1,2}(\\alpha)=\\frac{z_0\\left(-(2-\\alpha) \\pm  \\csqrt{\\alpha(2-\\alpha)}\\right)}{2}. \\]\n\\end{corollary}\n\n\n\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-2-double-totale.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-2-double-root-abs.pdf}\n\\caption{Paths \\(z(\\alpha)\\) of zeros of the fractional derivatives of \n\\(p(x)=x^2+(2+6i)x-12+9i\\) where \\(\\RL{0}{1/2}p\\) has a double root.  In the plot on the right the wide, light green graph represents the real part of \\(z(\\alpha\\)), while the thin, dark red graph represents its imaginary part. \n}\\label{fig double}\n\\end{figure}\n\n\nAnother natural question to ask is whether, given that a quadratic polynomial has  distinct zeros, can its fractional derivative have a double zero. Setting \\(z_1(\\alpha)=z_2(\\alpha)\\) one gets:\n\n\\begin{corollary}\\label{lem double zero}\nLet $p(x)=(x-a)^2+c_1(x-a)^1+c_0(x-a)^0$. Then the fractional derivative \\(\\RL{a}{\\alpha}(p(x))\\) will have a double zero precisely for one \\(\\alpha\\in\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}\\setminus\\N\\), namely: \n\\(\\alpha = 1-\\frac{c_1^2}{8c_0- c_1^2}\\).\n\\end{corollary}\n\n\n\n\n\\section{Flow of Zeros}\\label{sec flow}\n\nAs stated above, one of our main goals was to consider the paths of zeros of the fractional derivatives of polynomials \\(p(x) \\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) of arbitrary degrees. Unfortunately, unlike in the linear and quadratic cases, already for the cubics we find that the situation becomes considerably more complicated. This can be seen from the fact that one of the nicest properties -- the natural ordering of zeros -- fails already for degree 3: in other words, \nit is not true in general that zeros furthest away from the origin will yield the longest paths of zeros of fractional derivatives on its way to the origin. Figure\n\\ref{fig longest} shows a notable counterexample.\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-furthest-longest-counterexample-paths.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-furthest-longest-counterexample-abs.pdf}\n\\caption{The paths \\(z(\\alpha)\\) of zeros of the fractional derivatives of \nthe cubic \\(p(x)=(x-3-2i)(x-2+5i)(x-4+4i)\\)\nand the absolute values \\(|z(\\alpha)|.\\)\nIn the latter case, the zero of \\(p(x)\\) with the greatest absolute value is\nnot the starting point of the longest path.\n}\n\\label{fig longest}\n\\end{figure}\n\n\nHowever, certain convergence properties of the paths can be established in general. For example, the following theorem shows that all the paths terminate in the origin $a$. \n\n\\begin{theorem}\\label{theo roots lim}\nLet \\(p(x) \\in \\mathbb{C}[x]\\) of degree \\(n\\) such that for all \\(\\alpha\\in [0,n]\\)\nthe fractional derivative \\(p^{(\\alpha)}(x)\\) has no double zeros.\nThen there is an ordering of the roots \\(z_1,\\,z_2,\\,...,\\,z_n\\) or \\(p\\) such that for \n\\(p^{(\\alpha)}(z_j(\\alpha)) = 0\\) and \\(z_j(0)  = z_j\\) for \\(j\\in\\{1,...,n\\}\\) we have  \n\\(\\lim_{\\alpha \\to j} z_j(\\alpha) = a\\)\n\\end{theorem}\n\n\n\n\n\n\\begin{proof}\nWe  denote the coefficients in the expansion proved in Lemma \\ref{lem derv} by\n\\begin{equation}\nd^\\alpha_j:=\\prod_{k=j+1}^n\\!(k-\\alpha)\\cdot \\frac{j!}{n!}. \\label{eq dalphaj}\n\\end{equation}\nLet \\(0\\le j\\le n\\) and \\(m>j\\).  Here, clearly\n\\[\n\\lim_{\\alpha\\to m} \nd^\\alpha_j=\n\\lim_{\\alpha\\to m} \n\\prod_{k=j+1}^n\\!(k-\\alpha)\\cdot \\frac{j!}{n!}=0. \n\\]\nWrite the coefficients of the derivatives  \\(\\RL{a}{\\alpha}p(x)\\) by\n\\(d^\\alpha_j\\,c_j\\) \nas symmetric functions of the roots \n\\(z_1(\\alpha),z_2(\\alpha), \\dots, z_n(\\alpha)\\) of  \\(\\RL{a}{\\alpha}p(x)\\).\nBecause \n\\[\n0=\\lim_{\\alpha\\to m} d^\\alpha_0\\,c_0= \\lim_{\\alpha\\to m} \\prod_{k=1}^n z_k(\\alpha)\n\\]\nwe have\n\\(\\lim_{\\alpha\\to m} z_{k_1}(\\alpha)=0\\), for at least one \\(1\\le k_1\\le n\\).\nInductively, continuing with the next coefficient we get:\n\\begin{equation}\\label{eq sym 1}\n0=\\lim_{\\alpha\\to m} d^\\alpha_1\\,c_1= \\lim_{\\alpha\\to m} \n\\sum_{l=1}^n \\prod_{k\\ne l} z_k(\\alpha).\n\\end{equation}\nThere is one summand in (\\ref{eq sym 1}) that does not contain \\(z_{k_1}(\\alpha)\\).\nSo we need to have \\(\\lim_{\\alpha\\to m} z_{k_2}(\\alpha)=0\\) for at least one \\(k_2\\ne k_1\\).\nThis argument also holds for all \\(d^\\alpha_j\\,c_j\\) with \\(j<m\\).\nTherefore, we get \\(\\lim_{\\alpha\\to m} z_{k}(\\alpha)=0\\) for at least \\(m\\) distinct \n\\(k\\in\\{1,\\dots, n\\}\\).\n\\end{proof}\n\n\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-3-p1-totale.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-deg-3-p1-totale-mahler.pdf}\n\\caption{Paths \\(z(\\alpha)\\) of zeros of \\(\\RL{0}{\\alpha}(x^3+x^2+x+1+i)\\).\nillustrating the growth of the absolute value of \\(z(\\alpha)\\) as\n\\(\\alpha\\to\\infty\\) and \\(\\alpha\\to-\\infty\\) and the Mahler measure\nof \\(\\RL{0}{\\alpha}p\\) along with the bounds from Theorems \\ref{theo upper bound} and \\ref{theo lower bound}.\n}\n\\label{fig p1}\n\\end{figure}\n\n\\section{Bounds}\\label{sec bounds}\n\nAs stated in the introduction, for \\(f\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) the Gauss-Lucas theorem states that all zeros of \\(f'\\) lie in the convex hull of the set of zeros of \\(f\\), see \\cite[Theorem 6.1]{marden1966}. By induction this generalizes to all integral derivatives. Unfortunately, although all roots of the fractional derivatives converge to the origin, by our Theorem \\ref{theo roots lim}, the analogue of the Gauss-Lucas theorem does not hold for the fractional derivatives.\nThis is an immediate consequence of a result by Genchev and Sendov\n{\\cite{genchev-sendov}}, which is also stated as {\\cite[Theorem 2]{nikolov-sendov}}:\n\n\\begin{theorem}\nLet \\(L:\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\to\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) be a linear operator, such that \\(L(p)\\ne 0\\) implies that\nthe convex hull of the set of roots of \\(p\\) contains the roots of \\(L(p)\\). \nThen \\(L\\) is a linear functional or there are \\(c\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}\\setminus\\{0\\}\\) and \\(k\\in\\N\\) such that \\(L(p)=cp^{(k)}\\).\n\\end{theorem}\n\nFigure \\ref{fig longest} illustrates this result by giving a specific counterexample to the Gauss-Lucas property for the case of the Riemann-Liouville fractional derivatives. Nevertheless, it is possible to make some useful statements about how the absolute values of zeros\n\\(z_k(\\alpha)\\)\nof the fractional derivatives \\(\\RL{0}{\\alpha} f\\)\ndecrease as \\(\\alpha\\) increases in terms of the Mahler measure of \\(f\\).\n\nLet \\(p(x)=x^n+\\sum_{j=1}^{n-1} c_j x^j=\\prod_{j=1}^n (x-z_j)\\in\\mathbb{C}[x]\\).\nFor the Mahler measure \\(M(f)\\) \\cite{mahler} we have\n\\[\nM(p) = \\exp\\left(\\int_0^1 \\log(|f(e^{2\\pi i\\theta})|)\\, d\\theta \\right)= \\prod_{j=1}^n \\max\\{1,|z_j|\\}.\n\\]\nDenote the height of \\(f\\) by\n\\(\n\\nheight{p} = \\max\\{c_0,\\dots,c_n\\}\n\\)\nand the length of \\(f\\) by\n\\(\n\\nlength{p}=|c_0|+\\dots+|c_n|.\n\\)\nRecall that Mahler was able to prove the bounds \n\\begin{equation}\\label{eq mahler height}\n\\binom{n}{\\lfloor n/2 \\rfloor}^{-1} \\nheight{p} \\le M(p) \\le \\nheight{p}  \\csqrt{n+1}\n\\end{equation}\nand\n\\begin{equation}\\label{eq mahler length}\n2^{-n} \\nlength{p} \\le M(p) \\le  \\nlength{p}.\n\\end{equation}\nFor \\(\\ntwo{p} = \\left(\\sum_{j=1}^n |c_j|^2\\right)^\\frac{1}{2}\\) we have Landau's inequality\n\\cite{landau1905}\n\\begin{equation}\\label{eq landau}\nM(p)\\le ||p||_2.\n\\end{equation}\nWe generalize the definition of the Mahler measure to fractional derivatives.\nLet \\(Z(\\alpha)\\) be the set of zeros of \\(\\RL{0}{\\alpha}p\\).  We set\n\\[\nM\\left(\\RL{0}{\\alpha}p\\right)=\\prod_{z\\in Z(\\alpha)}\\max\\{1,|z|\\}.\n\\]\nand prove bounds similar to Equations \\ref{eq mahler height},\n\\ref{eq mahler length}, and \\ref{eq landau}\nfor the fractional cases. \nWe first estimate the coefficients \\(d^\\alpha_j\\) from Proposition \\ref{theo roots lim} and then use them to derive bounds for \\(M\\left(\\RL{0}{\\alpha}p\\right)\\).\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-mahler-paths.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-caputo-mahler-paths.pdf}\n\\caption{Comparing paths \\(z(\\alpha)\\) of zeros of the Riemann-Liouville (left) and Caputo (right) fractional derivatives \nof the  quintic \n\\(p(x)=(x+1-2i)(x-3-2i)(x-2+5i)(x-3+3i)(x+1+5i)\\), for \\(0\\le\\alpha\\le 5\\).\n}\n\\label{fig quintic paths}\n\\end{figure}\n\n\\begin{lemma}\\label{lem dj}\nLet \\(n\\in\\N\\) and \\(\\alpha\\in\\mathbb{R}} \\newcommand{\\Z}{\\mathbb{Z}\\setminus\\N\\).\nLet\n\\(d^\\alpha_j=\\left(\\prod_{k=j+1}^n\\!(k-\\alpha)\\right)\\cdot \\frac{j!}{n!}\n\\) where \\(0 \\le j\\le n-1\\).  Then\n\\begin{enumerate}\n\\item\\label{dj gt 0}\n\\(|d^\\alpha_j| \\le \\frac{n-\\alpha}{n}\\), for  \\(0<\\alpha< n\\).\n\\item\\label{dj lt}\n\\(|d^\\alpha_j| \\le \\frac{\\ff{|n-\\alpha|}{n}}{n!}\\), \nwhere \\(\\ff{|n-\\alpha|}{n}=\\prod_{k=0}^{n-1} |n-\\alpha-k|\\),\nfor \\(\\alpha<0\\) and \\(\\alpha>n\\).\n\\item\\label{dj lt 0 gt 2n}\n\\( |d^\\alpha_j|\\ge\\frac{n-\\alpha}{n}\\),  for \\(\\alpha<0\\)\nand \\(\\alpha>2n\\).\n\\item\\label{dj gt n}\n\\(|d^\\alpha_j|\\ge\\frac{\\alpha-n}{n} \\binom{n-1}{(n-1)/2}^{-1}\\), for \\(n<\\alpha\\le 2n\\).\n\\end{enumerate}\n\\end{lemma}\n\\begin{proof}\n\\begin{enumerate}\n\\item\nFor \\(0\\le\\alpha < j+1\\), we have\n\\[\n|d^\\alpha_j| = \\frac{n-\\alpha}{n} \\prod_{k=j+1}^{n-1}\\frac{k-\\alpha}{k} \\cdot \\frac{j!}{j!}< \\frac{n-\\alpha}{n}.\n\\]\nWhen \\(j+1 < \\alpha\\) set \\(h:=\\lfloor\\alpha\\rfloor\\).  We get\n\\begin{align*}\n|d^\\alpha_j| \n&=\n\\frac{n-\\alpha}{n}\\cdot\\frac{(n-1-\\alpha)\\cdots(h+1-\\alpha)\\cdot(\\alpha-j-1)\\cdots(\\alpha-h)\\cdot j!}{(n-1)!}\n\\le \\frac{n-\\alpha}{n}.\n\\end{align*}\n\\item For \\(\\alpha<0\\) and \\(\\alpha>n\\), we have\n\\begin{align*}\n|d^\\alpha_j| \n&=\n\\frac{|n-\\alpha|\\cdots|j+1-\\alpha|\\cdot j!}{n!}\\\\\n&=\n\\frac{|\\alpha-n|\\cdots|\\alpha-j-1|\\cdot j!}{n!}\\\\\n&\\le\n\\frac{|\\alpha-n|\\cdots|\\alpha-1|}{n!}=\\frac{|n-\\alpha|_n}{n!}.\n\\end{align*}\nFor \\(\\alpha<0\\) and \\(\\alpha>2n\\), we have\n\\begin{align*}\n|d^\\alpha_j| \n&=\\frac{n-\\alpha}{n}\\cdot\\frac{(n-1-\\alpha)\\cdots(j+1-\\alpha)\\cdot j!}{(n-1)!} \\\\\n&\\geq\\frac{n-\\alpha}{n}\\cdot\\frac{(n-1)\\cdots(j+1)\\cdot j!}{(n-1)!} \n=\\frac{n-\\alpha}{n}.\n\\end{align*}\n\\item\nFor \\(\\alpha>2n\\), we have\n\\begin{align*}\n|d^\\alpha_j| \n=\\frac{\\alpha-n}{n}\\cdot\\frac{(\\alpha-n+1)\\cdots(\\alpha-j-1)\\cdot j!}{(n-1)!} \\ge \\frac{\\alpha-n}{n}.\n\\end{align*}\n\\item\nFor \\(2n>\\alpha>n\\), we have\n\\begin{align*}\n|d^\\alpha_j| \n&=\\frac{(\\alpha-n)\\cdots(\\alpha-j-1)\\cdot j!}{n!}\\\\\n&\\ge\\frac{\\alpha-n}{n}\\cdot\\frac{1\\cdots(n-j-1)\\cdot j!}{(n-1)!}\\\\\n&\\ge\\frac{\\alpha-n}{n} \\binom{n-1}{(n-1)/2}^{-1}.\n\\end{align*}\n\\end{enumerate}\n\\end{proof}\n\n\\begin{figure}\n\\includegraphics[width=0.45\\textwidth]{paper-mahler-abs.pdf}\n\\includegraphics[width=0.45\\textwidth]{paper-caputo-mahler-abs.pdf}\n\\caption{Comparing the absolute value of \\(z(\\alpha)\\) of zeros of the Riemann-Liouville (left) and Caputo (right) fractional derivatives \nof the  quintic \n\\(p(x)=(x+1-2i)(x-3-2i)(x-2+5i)(x-3+3i)(x+1+5i)\\), for \\(0\\le\\alpha\\le 5\\).\n}\n\\label{fig quintic abs}\n\\end{figure}\n\n\nWe now show that \\(M\\left(\\RL{0}{\\alpha} p\\right)\\) \nis bounded from above for \\(0<\\alpha<\\deg p\\) where the bound linearly \ndecreases with \\(\\alpha\\).  Furthermore, in Theorem \\ref{theo lower bound},\nwe see that for \\(\\alpha<0\\) and \\(\\alpha>\\deg p\\) the Mahler measure\n\\(M\\left(\\RL{0}{\\alpha} p\\right)\\) increases at least linearlily with \\(\\alpha\\).\n\nIn our proofs  we use the following notation. We set\n\\begin{align}\\label{eq equivalent poly}\n\\RLs{0}{\\alpha}f(x) = \\frac{\\Gamma(n+1-\\alpha)}{n!}x^\\alpha\\cdot \\RL{0}{\\alpha}f(x).\n\\end{align}\nBecause \\(\\RLs{0}{\\alpha}f(x)\\)\nhas the same roots as \\(\\RL{0}{\\alpha}\\) except for the additional root \\(0\\).\nwe have \n\\begin{align*}\nM\\left(\\RLs{0}{\\alpha}f\\right)  \n= M\\left(\\RL{a}{\\alpha}f)\\right).\n\\end{align*}\n\n\\begin{theorem}\\label{theo upper bound}\nLet \\(f(x)\\in\\mathbb{C}[x]\\) be monic and let \\(0<\\alpha<n\\).  Then\n\\begin{enumerate}\n\\item $M(\\RL{0}{\\alpha} f)  \\le \\frac{n-\\alpha}{n}\\nlength{f}+1$\n\\item $M(\\RL{0}{\\alpha} f)  \\le  \\csqrt{n+1} \\max\\left\\{\\frac{n-\\alpha}{n} \\nheight{f},1\\right\\}$\n\\item $M(\\RL{0}{\\alpha} f)  \\le \\frac{n-\\alpha}{n} \\ntwo{f}+1$\n\\end{enumerate}\n\\end{theorem}\n\\begin{proof}\nWrite \\(f(x)=\\sum_{j=1}^n c_j x^j\\).\nWith the notation from Lemma \\ref{lem derv} we have\n\\begin{align*}\n\\RLs{0}{\\alpha}(x)\n= x^{n}+ \\sum_{j=0}^{n-1}\n\\left(\n\\prod_{k=j+1}^n\\!(k-\\alpha)\\right)\\cdot\n\\frac{j!}{n!} c_j x^{j} = x^n+\\sum_{j=0}^{n-1} d^{\\alpha}_j\\,c_j\\, x^j.\n\\end{align*}\nThis yields the following three bounds for the Mahler measure of \\(\\RL{a}{\\alpha} f \\).\n\\begin{enumerate}\n\\item\nWith Lemma \\ref{lem dj} (\\ref{dj gt 0}), we get\n\\begin{align*}\n\\nlength{\\RLs{0}{\\alpha}}\n&=|d_0^{(\\alpha)}c_0|+|d_1^{(\\alpha)}c_1|+\\dots+|d_{n-1}^{(\\alpha)}c_{n-1}|+1\\\\\n&\\le \\frac{n-\\alpha}{n}\\left(|c_0|+|c_1|+\\dots+|c_{n-1}|\\right)+1\\\\\n&\\le \\frac{n-\\alpha}{n} \\nlength{f}+1.\n\\end{align*}\nWith Mahler's (\\ref{eq mahler length}), we get\n\\[\nM\\left(\\RL{a}{\\alpha} f\\right) =\nM\\left(\\RLs{0}{\\alpha}\\right) \\le\n\\nlength{\\RLs{0}{\\alpha}} \\le \\frac{n-\\alpha}{n}\\nlength{f}+1.\n\\]\n\\item\nBy Lemma \\ref{lem dj} (\\ref{dj gt 0}), we have\n\\begin{align*}\n\\nheight{\\RLs{0}{\\alpha}}\n&=\\max\\{|d_0^{(\\alpha)}c_0|,|d_1^{(\\alpha)}c_1|,\\dots,|d_{n-1}^{(\\alpha)}c_{n-1}|,1\\}\\\\\n&\\le \\max\\left\\{\\frac{n-\\alpha}{n}\\max\\{|c_0|,|c_1|,\\dots,|c_{n-1}|\\},1\\right\\}\n\\\\\n&\\le \\frac{n-\\alpha}{n}\\nheight{f}+1.\n\\end{align*}\nWith Mahler's (\\ref{eq mahler height}),\nwe get\n\\[\nM\\left(\\RL{a}{\\alpha} f\\right) =\nM\\left(\\RLs{a}{\\alpha} f\\right) \\le\n \\csqrt{n+1} \\nheight{\\RLs{0}{\\alpha}} \\le  \\csqrt{n+1}  \\max\\left\\{\\frac{n-\\alpha}{n} \\nheight{f},1\\right\\}.\n\\]\n\\item\nBy Lemma \\ref{lem dj} (\\ref{dj gt 0}), we have\n\\begin{align*}\n\\ntwo{\\RLs{0}{\\alpha}}\n&= \\csqrt{(d_0^{(\\alpha)}c_0)^2+\\dots+(d_{n-1}^{(\\alpha)}c_{n-1})^2+1}\\\\\n&\\le\\frac{n-\\alpha}{n} \\csqrt{c_0^2+\\dots+c_{n-1}^2}+1\\\\\n&\\le\\frac{n-\\alpha}{n}\\ntwo{f}+1.\n\\end{align*}\nWith Landau's (\\ref{eq landau}), we get\n\\[\nM(\\RL{a}{\\alpha} f )=\nM\\left(\\RLs{0}{\\alpha}\\right) \\le \\ntwo{\\RLs{0}{\\alpha}} \\le\\frac{n-\\alpha}{n} ||f||_2+1.\n\\]\n\\end{enumerate}\n\\end{proof}\n\n\nNow we are able to prove:\n\n\n\\begin{theorem}\\label{theo lower bound}\nLet \\(f(x)\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) be monic of degree \\(n\\).  Then\n\\begin{enumerate}\n\\item\\label{tlb1}\\(M(\\RL{0}{\\alpha} f)\\ge 2^{-n}\\frac{n-\\alpha}{n}(\\nlength{f}-1)+1\\), for  \\(\\alpha<0\\) and \\(\\alpha>2n\\)\n\\item\\label{tlb2}\\(M(\\RL{0}{\\alpha} f)\\le \\frac{\\ff{|n-\\alpha|}{n}}{n!}\\nlength{f}+1\\), for  \\(\\alpha<0\\) and \\(\\alpha>n\\)\n\\item\\label{tlb3}\\(M(\\RL{0}{\\alpha} f)\\ge 2^{-n}\\binom{n-1}{\\lceil(n-1)/2\\rceil}^{-1}\\frac{\\alpha-n}{n}\\nlength{f}\\), for \\(n<\\alpha\\le 2n\\)\n\\end{enumerate}\n\\begin{proof}\n\\begin{enumerate}\n\\item[\\ref{tlb1}.]\nBy Lemma \\ref{lem dj} (\\ref{dj lt 0 gt 2n}), we have \n\\begin{align*}\n\\nlength{\\RLs{0}{\\alpha}f}\n&=|d_0^{(\\alpha)}c_0|+|d_1^{(\\alpha)}c_1|+\\dots+|d_{n-1}^{(\\alpha)}c_{n-1}|+1\\\\\n&\\ge \\frac{n-\\alpha}{n}\\left(|c_0|+|c_1|+\\dots+|c_{n-1}|+1-1\\right)+1\\\\\n&=\\frac{n-\\alpha}{n}(\\nlength{f}-1)+1.\n\\end{align*}\nWith Mahler's (\\ref{eq mahler length}), we get\n\\[\nM(\\RL{0}{\\alpha} f) =M\\left(\\RLs{0}{\\alpha}\\right) \\ge 2^{-n} \\nlength{\\RLs{0}{\\alpha}} \\ge 2^{-n}\\frac{n-\\alpha}{n}(\\nlength{f}-1)+1.\n\\]\n\\item[\\ref{tlb2}.]\nBy Lemma \\ref{lem dj} (\\ref{dj lt 0 gt 2n}), we have\n\\begin{align*}\n\\nlength{\\RLs{0}{\\alpha}f}\n&= \\csqrt{(d_0^{(\\alpha)}c_0)^2+\\dots+(d_{n-1}^{(\\alpha)}c_{n-1})^2+1}\\\\\n&\\le\\frac{\\ff{|n-\\alpha|}{n}}{n!}\\csqrt{c_0^2+\\dots+c_{n-1}^2}+1\\\\\n&\\le\\frac{\\ff{|n-\\alpha|}{n}}{n!}\\ntwo{f}+1.\n\\end{align*}\n\n\n\n\nWith Landau's (\\ref{eq landau}),\nwe get\n\\[\nM\\left(\\RL{a}{\\alpha} f\\right) =\nM\\left(\\RLs{a}{\\alpha} f\\right) \\le\n\\nlength{\\RLs{a}{\\alpha} f}\n\\le \\frac{\\ff{|n-\\alpha|}{n}}{n!}\\ntwo{f}+1.\n\\]\n\\item[\\ref{tlb3}.]\nBy Lemma \\ref{lem dj} (\\ref{dj gt n})\n\\begin{align*}\n\\nlength{\\RLs{0}{\\alpha}}\n&=|d_0^{(\\alpha)}c_0|+|d_1^{(\\alpha)}c_1|+\\dots+|d_{n-1}^{(\\alpha)}c_{n-1}|+1\\\\\n&\\ge \\frac{\\alpha-n}{n}\\binom{n-1}{\\lceil(n-1)/2\\rceil}^{-1}\\left(|c_0|+|c_1|+\\dots+|c_{n-1}|+1\\right)\\\\\n&=\\frac{\\alpha-n}{n}\\binom{n-1}{\\lceil(n-1)/2\\rceil}^{-1}\\nlength{f}\n\\end{align*}\nWith Mahler's (\\ref{eq mahler length}), we get\n\\[\nM(\\RL{0}{\\alpha} f) =M\\left(\\RLs{0}{\\alpha}\\right) \\ge 2^{-n} \\nlength{\\RLs{0}{\\alpha}} \\ge 2^{-n}\\frac{\\alpha-n}{n}\\binom{n-1}{(n-1)/2}^{-1}\\nlength{f}.\n\\]\n\\end{enumerate} \n\\end{proof}\n\\end{theorem}\n\n\nIn Figure \\ref{fig p1} we present the paths of zeros and the\nMahler measures of the fractional derivatives of a degree 3 polynomial along with the bounds from Theorems \\ref{theo lower bound} and \\ref{theo upper bound}.\nFurthermore Figures \\ref{fig one}, \\ref{fig double}, and \\ref{fig p1} show \nthe growth of $M(\\RL{0}{\\alpha} p)$ for \\(\\alpha<0\\) and \\(\\alpha>0\\).\n\n\\begin{figure}\n\\begin{center}\n\\includegraphics[width=0.45\\textwidth]{paper-caputo-rieliou-mahler-bound.pdf}\n\\end{center}\n\\caption{Mahler measure of the Riemann-Liouville and Caputo fractional derivatives of the quintic \\(p(x)=(x+1-2i)(x-3-2i)(x-2+5i)(x-3+3i)(x+1+5i)\\)  for \\(0\\le\\alpha\\le 5\\) along with the bound from\nTheorem \\ref{theo upper bound} and Corollary \\ref{cor upper bound caputo}.}\n\\label{fig bound mahler}\n\\end{figure}\n\n\n\\section{Caputo Fractional Derivatives} \\label{caputo}\n\n \nSwitching the order of differentiation and integration in Definition \\ref{def rl} yields the Caputo fractional derivative, see \\cite{diethelm2010} for example.\n\n\\begin{definition}[Caputo fractional derivative]\\label{def caputo}\nLet \\(f\\) be analytic on a convex open set \\(C\\) let and \\(a\\in C\\).  Let $\\alpha>0$ and $m=\\lceil \\alpha \\rceil$ then the \\(\\alpha\\)-th Caputo fractional derivative is\n\\[\n\\CA{a}{\\alpha} f(t)\n=\n\\RLI{a}{m-\\alpha} \n\\frac{d^m}{dt^m} \nf(t).\n\\]\n\\end{definition}\nThe Caputo fractional derivative has the advantage that the composition rule holds, while the power rule is very similar to that of the Riemann-Liouville fractional derivative: \n\\[\n    \\CA{a}{\\alpha}x^\\beta= \\begin{cases}\n        0 & \n        \\begin{array}{l}\\text{if } \\beta \\in \\{0,1,\\dots, m-1\\}\n        \\end{array}\\\\\n        \\dfrac{\\Gamma(\\beta+1)}{\\Gamma(\\beta-\\alpha+1)}(x-a)^{\\beta-\\alpha} & \n\\begin{array}{l}        \n        \\text{if } \\beta\\in\\N \\text{ and }\\beta \\ge  m\\text{ or }\\\\[-0.5ex]\n        \\text{if } \\beta\\not\\in\\N \\text{ and }\\beta > m-1,\n\\end{array}\n    \\end{cases}\n\\]\nwhere $m=\\lceil \\alpha \\rceil$.\n\nThese derivatives differ, but obey some common general trends. Figures \\ref{fig quintic paths} and \\ref{fig quintic abs} compare the paths of zeros of the Riemann-Liouville and Caputo derivatives of a degree 5 polynomial.\n\nThe upper bounds from Theorem \\ref{theo upper bound} easily transfer \nto the Caputo fractional derivative.  Because the coefficients of the Caputo fractional derivatives of a polynomial \\(p\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) are either the \nsame as those of the Riemann-Liouville fractional derivative or zero, \nwe have\n\\[\n\\norm{\\CAs{0}{\\alpha}}{k} \\le \\norm{\\RLs{0}{\\alpha}}{k},\n\\]\nfor  \\(k\\in\\{1,2,\\infty\\}\\).  This yields the bounds:\n\\begin{corollary}\n\\label{cor upper bound caputo}\nLet \\(p\\in\\mathbb{C}} \\newcommand{\\N}{\\mathbb{N}[x]\\) be monic of degree $n$ and let \\(0<\\alpha<n\\).  Then \n\\begin{enumerate}\n\\item $M(\\CA{0}{\\alpha} p)  \\le \\frac{n-\\alpha}{n}\\nlength{p}+1$,\n\\item $M(\\CA{0}{\\alpha} p)  \\le  \\csqrt{n+1} \\max\\left\\{\\frac{n-\\alpha}{n} \\nheight{p},1\\right\\}$,\n\\item $M(\\CA{0}{\\alpha} p)  \\le \\frac{n-\\alpha}{n} \\ntwo{p}+1$.\n\\end{enumerate}\n\\end{corollary}\n\n\n\n\n\n\\section{Conclusion} \\label{conclusion}\n\nThe bounds we have established in Section 5 and Section 6 were sufficient for our purposes, but they are \nfar from best possible. \nFigure \\ref{fig bound mahler} illustrates the decline of $M(\\RL{0}{\\alpha} p)$ and $M(\\CA{0}{\\alpha} f)$, when $\\alpha$ approaches \\(n\\), as described by Theorem \\ref{theo upper bound} and Corollary \\ref{cor upper bound caputo} and Theorem \\ref{theo roots lim}. In a future work, it would be interesting to consider the true growth of the paths of zeros, for \\(\\alpha<0\\) and \\(\\alpha>n\\). Also, the maximal extent of loops of paths, after traversing the origin \\(a\\), is something that could be worth looking at. Moreover, as Figure \\ref{fig p1} clearly shows, the paths exhibit very distinct linear asymptotes in both directions \\(\\alpha\\to\\infty\\) and \\(\\alpha\\to-\\infty\\). For polynomials of higher degrees, their exact directions are not yet known. \n\nIn addition to this, as we have noted earlier, the useful Gauss-Lucas property does not hold universally for the fractional derivatives of polynomials. However, some of the dynamical properties we have observed could be investigated with insights related to those that play a key role in the integral case. In particular, Gauss himself suggested a very intriguing physical interpretation of the nontrivial critical points of a polynomial (the critical points which are not zeros) as the equilibrium points in certain force fields, generated by particles placed at the zeros of the polynomial, with masses equal to the multiplicity of the zeros and repelling with a force inversely proportional to the distance. This amazing physical application of a purely theoretical polynomial concept is exceedingly intriguing and should be investigated further. It could go a long way in explaining the profound intricacies of the paths of zeros, and their seemingly chaotic local behavior.  \n\n\n\n\n\n\n\n", "meta": {"timestamp": "2023-01-20T02:15:14", "yymm": "2301", "arxiv_id": "2301.08149", "language": "en", "url": "https://arxiv.org/abs/2301.08149", "abstract": "We investigate the behavior of fractional derivatives of polynomials. In particular, we consider the locations and the asymptotic behaviour of their zeros and give bounds for their Mahler measure.", "subjects": "General Mathematics (math.GM)", "title": "Zeros of fractional derivatives of polynomials", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9927672366245092, "lm_q2_score": 0.8539127492339909, "lm_q1q2_score": 0.8477366003754666}}
{"url": "https://arxiv.org/abs/1701.05855", "title": "Judicious partitions of uniform hypergraphs", "abstract": "The vertices of any graph with $m$ edges may be partitioned into two parts so that each part meets at least $\\frac{2m}{3}$ edges. Bollobás and Thomason conjectured that the vertices of any $r$-uniform hypergraph with $m$ edges may likewise be partitioned into $r$ classes such that each part meets at least $\\frac{r}{2r-1}m$ edges. In this paper we prove the weaker statement that, for each $r\\ge 4$, a partition into $r$ classes may be found in which each class meets at least $\\frac{r}{3r-4}m$ edges, a substantial improvement on previous bounds.", "text": "\\section{Introduction}\nThe vertices of any graph (indeed, any multigraph) may be partitioned into two parts, each of which meets at \nmost two thirds of the edges (\\cite{BS93}; it also appears as a problem in \\cite{MGT}). An equivalent statement in \nthis case is that each part spans at most one third of the edges. These two statements give rise to different \ngeneralisations when a partition into more than two parts is considered. In this paper we shall only address \nthe problem of meeting many edges; the problem of spanning few edges is addressed in \\cite{BS99} for the \ngraph case and \\cite{BS97} for the hypergraph case.\n\nA particularly interesting case occurs when we partition the vertices of an $r$-uniform hypergraph into $r$ \nclasses. Bollob\\'as and Thomason (see \\cite{BRT93}, \\cite{BS00}) conjectured that every $r$-uniform hypergraph \nwith $m$ edges has an $r$-partition in which each class meets at least $\\frac{r}{2r-1}m$ edges.\nThe author \\cite{Hasa} recently proved the conjecture for the case $r=3$. \n\nThe previous best known bound for each $r>3$ was proved by Bollob\\'as and Scott \\cite{BS00}, who in fact \nobtained a constant independent of $r$. The method they used was to progressively refine a partition by \nrepartitioning the vertices in two or three parts to increase the number of parts which met $cm$ edges; in \ndoing so they showed that $r$ disjoint parts meeting at least $cm$ edges may be found for $c=0.27$.\n\nOur strategy will be to combine those ideas with the methods used to prove the conjectured bound in the case \n$r=3$ \\cite{Hasa}. We shall obtain values of $c_r=\\frac{r}{3r-4}$ for $r>3$, and so $c_r \\to \\frac{1}{3}$ as $r \\to \\infty$. While this is a significant improvement on previous bounds, it is still some distance from the \nconjectured bounds which approach $\\frac{1}{2}$ in the limit.\n\nAll results obtained in this paper apply to hypergraphs in which repeated edges are permitted, and in \nfact we shall need this extra generality in an induction step.\n\n\\section{New parts from old}\nIn this section we give two lemmas which show that if we have a partition in which some parts meet many more \nedges than required we may locally refine that partition and increase the number of parts which meet the \nrequired number of edges. The first of these lemmas was proved in \\cite{BS00}; the second is a new result in \nthe same spirit. The setting for each lemma is the same. $G$ is a multi-hypergraph, but not necessarily \nuniform; edges may have any number of vertices, and edges of any size may be repeated. $G$ has $m$ edges, but \nthe maximum degree of a vertex is less than $cm$ where $c>0$ is an arbitrary constant. For $A,B\\subset V$\nwe write $d(A)$ for the number of edges meeting $A$ and $d(A,B)$ for the number meeting both $A$ and $B$ (we \nshall only use the latter notation where $A$ and $B$ are disjoint).\n\n\\begin{lemma}[\\cite{BS00}]\\label{rutwothree}\nLet $c>0$ be a constant and let $G$ be a multi-hypergraph on vertex set $V$ with $m$ edges such that $\\Delta(G)<cm$. If $A$ and $B$ are disjoint subsets of $V$, each of which meet at least $2cm$ edges, then there is a partition of $A\\cup B$ into three parts, each of which meets at least $cm$ edges.\n\\end{lemma}\nLemma \\ref{rutwothree} was proved in \\cite{BS00}; we provide their proof for completeness.\n\\begin{proof}\nWe may replace each edge by a subedge if necessary so that no edge meets either $A$ or $B$ in more than \none vertex. We may then partition $A$ into three parts, $A_1, A_2, A_3$, such that the union of any two\nmeets at least $cm$ edges: we may do this by taking $A_1$ to be a maximal part meeting fewer than $cm$ edges,\nand then dividing $A\\setminus A_1$ into two non-empty parts, since any set which strictly contains $A_1$ must \nmeet at least $cm$ edges and $A_2\\cup A_3=A\\setminus A_1$ meets $d(A)-d(A_1)>2cm-cm=cm$ edges. We find a \nsimilar partition $B=B_1\\cup B_2\\cup B_3$. It is sufficient to find $i, j$ such that $d(A_i\\cup B_j)\\ge cm$; \nthen $A_i\\cup B_j, A\\setminus A_i, B\\setminus B_j$ will be a suitable partition. We shall show that this is \nalways possible.\n\nSince each edge meets $A$ in at most one place then $d(A)=\\sum_id(A_i)$ (and likewise $d(B)=\\sum_id(B_i)$). \nSimilarly, if an edge meets both $A$ and $B$ then we may find unique $i, j$ for which it meets $A_i$ and $B_j$. \nTherefore,\n\\begin{equation*}\nd(A,B)=\\sum_{i,j}d(A_i,B_j).\n\\end{equation*}\nAlso,\n\\begin{equation*}\nd(A_i\\cup B_j)=d(A_i)+d(B_j)-d(A_i,B_j),\n\\end{equation*}\nso\n\\begin{eqnarray*}\n\\sum_{i,j}d(A_i\\cup B_j) &=& \\sum_{i,j}\\left(d(A_i)+d(B_j)-d(A_i,B_j)\\right)\\\\\n &=& 3d(A)+3d(B)-\\sum_{i,j}d(A_i,B_j)\\\\\n &=& 3d(A)+3d(B)-d(A,B)\\\\\n &\\ge& 3d(A)+2d(B)\\ge 10cm.\n\\end{eqnarray*}\nSince there are nine terms in the LHS, at least one must exceed $cm$. \n\\end{proof}\n\nThis proof shows more than required: we can always find a partition of $A\\cup B$ into three parts, two meeting at least $cm$ edges and the third meeting at least $\\frac{10cm}{9}$. However, we cannot always find a partition into three parts, two meeting more than $cm+1$ edges and the third meeting at least $cm$, as seen by considering the case where each of $A$ and $B$ have three vertices, two meeting $cm-1$ edges and one meeting 2 edges. \n\nA part which does not meet the desired number of edges can also be useful provided we have another part meeting sufficiently many edges to combine with it.\n\n\\begin{lemma}\\label{rugoodbad}\nLet $c>0$ be a constant and let $G$ be a multi-hypergraph on vertex set $V$ with $m$ edges such that $\\Delta(G)<cm$. If $A$ and $B$ are disjoint subsets of $V$, with $d(A)\\ge 2cm$ and $d(A)+2d(B)\\ge 3cm$, then there is a partition of $A\\cup B$ into two parts, each of which meets at least $cm$ edges.\n\\end{lemma}\n\\begin{proof}\nAs before, we may replace each edge by a subedge if necessary so that no edge meets either $A$ or $B$ in \nmore than one vertex. We may then again partition $A$ into three parts $A_1, A_2, A_3$ such that the union \nof any two meets at least $cm$ edges. It is now sufficient to find $i$ such that $d(A_i\\cup B)\\ge cm$; then \n$A_i\\cup B, A\\setminus A_i$ will be a suitable partition. We claim this is always possible.\n\nAgain,\n\\begin{equation*}\nd(A,B)=\\sum_{i}d(A_i,B)\n\\end{equation*}\nand\n\\begin{equation*}\nd(A_i\\cup B)=d(A_i)+d(B)-d(A_i,B),\n\\end{equation*}\nso\n\\begin{eqnarray*}\n\\sum_{i}d(A_i\\cup B) &=& \\sum_{i}\\left(d(A_i)+d(B)-d(A_i,B)\\right)\\\\\n &=& d(A)+3d(B)-\\sum_{i}d(A_i,B)\\\\\n &=& d(A)+3d(B)-d(A,B)\\\\\n &\\ge& d(A)+2d(B)\\ge 3cm.\n\\end{eqnarray*}\nSince there are three terms in the LHS, at least one must be at least $cm$. \n\\end{proof}\nIn particular, we may find such a partition when $d(A)\\ge 2cm$ and $d(B)\\ge \\frac{cm}{2}$; \nthis is the only case in which we shall apply Lemma \\ref{rugoodbad}.\n\n\\section{The bounds}\nThroughout this section $G$ is an $r$-uniform multi-hypergraph on vertex set $V$ with $m$ edges. \nOur goal is to prove that there is a partition of $V$ into $r$ parts such that each part meets at least \n$c_r m$ edges for some suitable constant $c_r$. In order to use Lemmas \\ref{rugoodbad} and \n\\ref{rutwothree} we need to reduce to the case $\\Delta(G)<c_rm$. To that end we apply induction on $r$. If \nsome vertex meets at least $c_r m$ edges then we may remove that vertex, replacing each edge of $G$ by a subedge of \nsize $r-1$ not containing that vertex, and then use the result for $r-1$ to partition the remaining vertices \ninto $r-1$ parts, each meeting at least $c_{r-1} m$ edges; this is sufficient so long as $c_{r-1} \\ge c_r$, \nwhich will be the case. We shall take $c_2=\\frac{2}{3}$, so the case $r=2$ is known.\n\nOur plan will be to look at partitions for which $\\sum_id(V_i)$ is large, and show that if some parts meet too \nfew edges then there are enough parts meeting at least $2c_rm$ edges to allow us to construct a good partition \nby combining parts as above. \n\nCertainly any partition which is optimal in the sense of maximising $\\sum_id(V_i)$ also satisfies the \nlocal optimality condition that $\\sum_id(V_i)$ cannot be increased by moving a single vertex. We shall\nconsider particularly the yet weaker condition that $\\sum_id(V_i)$ cannot be increased by moving a single\nvertex into $V_r$. We begin by establishing bounds on partitions satisfying this condition.\n\n\\begin{lemma}\\label{aaa}\nLet $V_1, V_2, \\ldots, V_r$ be a partition for which $\\sum_i d(V_i)$ cannot be increased by\nmoving a vertex into $V_r$. Then \n\\begin{equation*}\n\\sum_{i=1}^r d(V_i) \\ge (r+1)m-rd(V_r).\n\\end{equation*}\n\\end{lemma}\n\\begin{proof}\nFor each $i<r$ and each $v\\in V_i$, since moving $v$ into $V_r$ does not increase the sum, the number of edges\n$e$ such that $e\\cap V_i = \\{v\\}$ must be at least the number of edges $e$ containing $v$ which do not meet \n$V_r$: the first quantity is the decrease in $d(V_i)$ effected by moving $v$ and the second is the increase \nin $d(V_r)$. Thus\n\\begin{equation*}\n\\sum_{i\\ne r}\\sum_{v\\in V_i}|\\{e:e\\cap V_i={v}\\}|\\le \\sum_{i\\ne r}\\sum_{v\\in V_i}|\\{e\\ni v: e\\cap V_r =\\emptyset\\}|.\n\\end{equation*}\nSince, for $v\\ne w$, $\\{e:e\\cap V_i={v}\\}$ and $\\{e:e\\cap V_i={w}\\}$ are disjoint,\n\\begin{eqnarray*}\n\\sum_{i\\ne r}\\sum_{v\\in V_i}|\\{e:e\\cap V_i={v}\\}| &=& \\sum_{i\\ne r}|\\bigcup_{v\\in V_i}\\{e:e\\cap V_i={v}\\}| \\\\\n &=& \\sum_{i\\ne r}|\\{e:|e\\cap V_i|=1\\}|.\n\\end{eqnarray*}\nFor $1\\le i <r$, let $E_i=\\{e:|e\\cap V_i|=1\\}$. For each edge $e$ write $f(e)$ for the number of parts (including $V_r$)\nwhich meet $e$. If $f(e)<r$ then $|e\\cap V_i|>1$ for some $i$, and so $e$ is in at most $f(e)-1$ of the $E_i$; \ntrivially if $f(e)=r$ then $e$ is in at most $r-1=f(e)-1$ of the $E_i$. Thus \n\\begin{eqnarray*}\n\\sum_{i\\ne r}|\\{e:|e\\cap V_i|=1\\}| &\\le& \\sum_{e}(f(e)-1) \\\\\n &=& \\sum_{i=1}^r d(V_i)-m.\n\\end{eqnarray*}\nAlso,\n\\begin{equation*}\n\\sum_{i\\ne r}\\sum_{v\\in V_i}|\\{e\\ni v: e\\cap V_r =\\emptyset\\}|=r(m-d(V_r)),\n\\end{equation*}\nsince each edge not meeting $V_r$ is counted exactly $r$ times in the sum, once for each vertex it contains. \nCombining the above relations, we see that\n\\begin{equation*}\n\\sum_{i=1}^r d(V_i)-m \\le r(m-d(V_r)),\n\\end{equation*}\nas required.\n\\end{proof}\n\nThe reason for considering the condition that $\\sum_id(V_i)$ cannot be increased by moving a single\nvertex into $V_r$ is that, as we shall see, it is preserved by moving vertices into $V_r$. If we are able \nto start from a partition which satisfies the condition and in which $V_1,\\ldots, V_{r-1}$ are ``good'' (in \nthe sense of meeting at least a certain proportion of edges) then we can try to improve the partition by \nmoving vertices into $V_r$ while keeping the other parts good. In this way we will either obtain a partition\ninto $r$ good parts or we will be forced to stop because $V_1,\\ldots, V_{r-1}$ are all minimal good sets. We formalise these ideas in the following lemma.\n\n\\begin{lemma}\\label{aab}\nLet $V_1, V_2, \\ldots, V_r$ be a partition for which $\\sum_i d(V_i)$ is as large as possible, ordered such that\n$d(V_1)\\ge d(V_2)\\ge \\cdots \\ge d(V_r)$, and let $c>0$ be a constant. If $d(V_{r-1})\\ge cm$ then either there \nexists a partition $W_1, W_2, \\ldots, W_r$, with $W_i\\subseteq V_i$ for $1\\le i\\le r-1$ \n(and so $W_r\\supseteq V_r$), such that each part meets at least $cm$ edges, or there exists a partition \n$W_1, W_2, \\ldots, W_r$, again with $W_i\\subseteq V_i$ for $1\\le i\\le r-1$,\nsuch that, for each $i\\ne r$, $d(W_i)\\ge cm$ but $d(W_i\\setminus \\{w\\})< cm$ for any $w\\in W_i$, and also\n\\begin{equation*}\n\\sum_{i=1}^{r-1} d(W_i) > (r+1)(m-cm).\n\\end{equation*}\n\\end{lemma}\n\\begin{proof}\nSuppose $U_1, U_2, \\ldots, U_r$ is a partition with $U_i\\subseteq V_i$ for each $i<r$ and so $U_r\\supseteq V_r$. \nFor any $v\\notin U_r$, say $v\\in U_i$, let $U'_i=U_i\\setminus\\{v\\}$, $U'_r=U_r\\cup\\{v\\}$, $V'_i=V_i\\setminus\\{v\\}$ \nand $V'_r=V_r\\cup\\{v\\}$. Then\n\\begin{eqnarray*}\nd(U_i)-d(U'_i)&=&|\\{e:e\\cap U_i=\\{v\\}\\}| \\\\\n &\\ge&|\\{e:e\\cap V_i=\\{v\\}\\}| \\\\\n &=&d(V_i)-d(V'_i)\n\\end{eqnarray*}\nand\n\\begin{eqnarray*}\nd(U'_r)-d(U_r)&=&|\\{e\\ni v:e\\cap U_r=\\emptyset\\}| \\\\\n &\\le&|\\{e\\ni v:e\\cap V_r=\\emptyset\\}| \\\\\n &=&d(V'_r)-d(V_r)\n\\end{eqnarray*}\nso\n\\begin{eqnarray*}\nd(U'_i)+d(U'_r) &\\le& \\left(d(U_i)+d(V'_i)-d(V_i)\\right)+\\left(d(V'_r)-d(V_r)+d(U_r)\\right) \\\\\n &=&d(U_i)+d(U_r)+\\left(d(V'_i)+d(V'_r)-d(V_i)-d(V_r)\\right) \\\\\n &\\le&d(U_i)+d(U_r),\n\\end{eqnarray*}\nmeaning that $U_1, U_2, \\ldots, U_r$ also satisfies the condition that $\\sum_i d(U_i)$ cannot be increased by moving a vertex into $U_r$.\n\nFor each $i<r$, then, let $W_i$ be a minimal subset of $V_i$ satisfying $d(W_i)\\ge cm$, and let $W_r=V\\setminus\\bigcup_{i<r}W_i$. \nIf $d(W_r)\\ge cm$ then this is a suitable partition with each part meeting at least $cm$ edges; if not then Lemma \\ref{aaa} ensures that\n\\begin{eqnarray*}\n\\sum_{i=1}^{r-1} d(W_i) &\\ge& (r+1)m-rd(W_r)-d(W_r) \\\\\n &>&(r+1)(m-cm),\n\\end{eqnarray*}\nas required. \n\\end{proof}\n\nWe are now ready to prove the main result. We shall show that if we start from a partition maximising $\\sum_id(V_i)$ then either we have enough elbow room to obtain a good partition by repeated application of Lemmas \\ref{rutwothree} and \\ref{rugoodbad} or we may use Lemma \\ref{aab} to obtain a good partition.\n\n\\begin{theorem}\\label{rumain}\nLet $c_2=\\frac{2}{3}$, $c_3=\\frac{5}{9}$ and $c_r=\\frac{r}{3r-4}$ for $r>3$. If $G$ is an $r$-uniform \nmulti-hypergraph with $m$ edges there is a partition of the vertex set into $r$ parts with each part meeting at \nleast $c_r m$ edges.\n\\end{theorem}\n\\begin{proof}\nWe use induction on $r$; the case $r=2$ is known. For $r>2$, if some vertex $v$ meets at least \n$c_r m$ edges then we may, by replacing each edge with a subedge of size $r-1$ not containing $v$ and applying \nthe $r-1$ case, find a partition of the other vertices into $r-1$ parts each meeting at least $c_{r-1}m > c_rm$ \nedges. Together with $\\{v\\}$, this is a suitable $r$-partition. Thus we may assume that no vertex meets $c_r m$ \nedges.\n\nNow let $V_1, V_2, \\ldots, V_r$ be a partition for which $\\sum_i d(V_i)$ is as large as possible, ordered such \nthat $d(V_1)\\ge d(V_2)\\ge \\cdots \\ge d(V_r)$. If $d(V_r)\\ge c_r m$ then we are done, so we may assume \n$d(V_r)< c_r m$. We consider three cases based on the values of $d(V_{r-1})$ and $d(V_r)$.\n\n\\vspace{5pt}\n\\textbf{Case 1.} $d(V_{r-1})\\ge c_r m$.\n\nBy Lemma \\ref{aab}, either we have a good partition or we may find a partition $W_1, W_2, \\ldots, W_r$ such that, for each\n$i<r$, $d(W_i)\\ge c_r m$ but $W_i$ is minimal for this property, and further that $\\sum_{i=1}^{r-1} d(W_i) \\ge (r+1)(m-c_r m)$.\nSuppose the partition found is of the latter type. For each $i<r$, since $\\Delta(G)< c_r m$, $|W_i|>1$. Let $v, w$ be two \nvertices in $W_i$; by minimality of $W_i$ the number of edges meeting $W_i$ only at $v$ is more than $d(W_i)-c_r m$, and so \nis the number meeting $W_i$ only at $w$. These sets of edges are disjoint from each other and from the set of edges \nmeeting $W_i$ in more than one vertex; thus, writing $d_2(X)$ for the number of edges meeting $X$ in more than one vertex,\n\\begin{equation*}\nd(W_i) > 2(d(W_i)-c_r m)+d_2(W_i),\n\\end{equation*}\ni.e.\n\\begin{equation*}\nd(W_i)+d_2(W_i)<2c_r m.\n\\end{equation*}\nHowever, each edge not meeting $W_r$ meets at least one other part at more than one vertex, so\n\\begin{equation*}\n\\sum_{i=1}^{r-1}d_2(W_i)>m-c_r m.\n\\end{equation*}\nCombining this with the bound on $\\sum_{i=1}^{r-1} d(W_i)$ gives\n\\begin{equation*}\n\\sum_{i=1}^{r-1}(d(W_i)+d_2(W_i))>(r+2)(m-c_r m),\n\\end{equation*}\nso for some $i<r$\n\\begin{equation*}\nd(W_i)+d_2(W_i)>\\frac{r+2}{r-1}(m-c_r m).\n\\end{equation*}\nConsequently, using our upper bound on $d(W_i)+d_2(W_i)$,\n\\begin{eqnarray*}\n&&\\frac{r+2}{r-1}(m-c_r m) < 2c_r m \\\\\n&\\Rightarrow&(r+2)(1-c_r) < 2c_r(r-1) \\\\\n&\\Rightarrow& r+2 < 3rc_r\n\\end{eqnarray*}\nand so $c_r > \\frac{r+2}{3r}$. \n\nFor $r=3$, $c_r=\\frac{5}{9}=\\frac{r+2}{3r}$; for $r\\ge 4$,\n\\begin{eqnarray*}\nc_r &=& \\frac{r}{3r-4} \\\\\n&=& \\frac{r+2}{3r}-\\frac{2-8r}{3r(3r-4)} \\\\\n&\\le& \\frac{r+2}{3r};\n\\end{eqnarray*}\nin either case a contradiction is obtained (from our assumption that we did not get a good partition from Lemma \\ref{aab}). (End of Case 1.)\n\\vspace{1.5ex}\n\nNote that, using Lemma \\ref{aaa}, if $d(V_r) < c_r m$ then \n\\begin{eqnarray*}\nd(V_{r-1})-c_r m&=&\\sum_{i=1}^{r} d(V_i)-\\sum_{i=1}^{r-2} d(V_i)-d(V_r)-c_rm \\\\\n&>& \\sum_{i=1}^{r} d(V_i)-(r-2)m-2c_r m\\\\\n&\\ge& (r+1)m-rd(V_r)-(r-2)m -2c_r m\\\\\n&>& (r+1)m-rc_rm-(r-2)m -2c_r m\\\\\n&=& 3m-(r+2)c_r m,\n\\end{eqnarray*}\nso for $r=3$, since $c_r < \\frac{3}{r+2}$, Case 1 is the only possible case. For the remaining cases, then, we assume $r\\ge 4$.\n\n\\vspace{5pt}\n\\textbf{Case 2.} $c_r m > d(V_{r-1}) \\ge d(V_r) \\ge \\frac{c_rm}{2}$.\n\nSuppose that $k$ parts meet at least $2c_r m$ edges, $l$ meet fewer than $c_r m$, and the remaining $r-k-l$ meet at least $c_r m$\nbut fewer than $2c_r m$. Using Lemma \\ref{rugoodbad} we may combine a part meeting at least $2c_r m$ edges and a part meeting at least\n$\\frac{c_rm}{2}$ to produce two parts meeting at least $c_r m$; we know, since $V_r$ meets fewest edges, that each part meets at\nleast $\\frac{c_rm}{2}$ and so we can obtain a good partition provided $k\\ge l$. We shall show that this must be so.\n\nSince $r \\ge 4$, $2c_r\\le 1$, and since $c_r m > d(V_{r-1})$, $l\\ge 2$. So\n\\begin{eqnarray*}\n\\sum_{i=1}^rd(V_i) &<& 2c_r m(r-k-l)+c_r ml+mk \\\\\n&=& 2c_r mr -c_r lm + (1-2c_r)km.\n\\end{eqnarray*}\nHowever, by Lemma \\ref{aaa},\n\\begin{equation*}\n\\sum_{i=1}^r d(V_i) > (r+1)m-rc_r m,\n\\end{equation*}\nso\n\\begin{equation*}\n2rc_r -lc_r + k(1-2c_r) > (r+1)-rc_r;\n\\end{equation*}\nhowever, if $k<l$ then\n\\begin{eqnarray*}\n2rc_r -lc_r + k(1-2c_r) &\\le& 2rc_r-lc_r+(l-1)(1-2c_r) \\\\\n&=& (2r-1)c_r+(l-1)(1-3c_r)\\,.\n\\end{eqnarray*}\nSince $l\\ge 2$ and $1-3c_r<0$,\n\\begin{eqnarray*}\n(2r-1)c_r+(l-1)(1-3c_r)&\\le& (2r-4)c_r+1 \\\\\n&=& r+1-rc_r\n\\end{eqnarray*}\n(the final equality follows since $c_r=\\frac{r}{3r-4}$), and a contradiction is obtained.\n\n\\vspace{5pt}\n\\textbf{Case 3.} $c_r m > d(V_{r-1})$ and $\\frac{c_rm}{2} > d(V_r)$.\n\nSuppose that $j$ parts meet at least $2c_r m$ edges, $k$ meet at least $\\frac{c_rm}{2}$ but fewer than $c_r m$, \n$l$ meet fewer than $\\frac{c_rm}{2}$, and the remaining $r-j-k-l$ meet at least $c_r m$ but fewer than $2c_r m$. \nUsing Lemma \\ref{rugoodbad} $k$ times and Lemma \\ref{rutwothree} $l$ times, we may find $r$ disjoint parts which meet at least $c_r m$ edges\nprovided that $j\\ge k+2l$ (each application of Lemma \\ref{rugoodbad} requires one part which meets $2c_r m$ edges and each application of Lemma \\ref{rutwothree} requires two). We shall show that this must be so.\n\nFrom Lemma \\ref{aaa}, since $d(V_r) < \\frac{c_rm}{2}$ and $r\\ge 4$, \n\\begin{eqnarray*}\n\\frac{1}{r}\\sum_{i=1}^r \\frac{d(V_i)}{m} &>& \\frac{r+1}{r}-\\frac{c_r}{2} \\\\\n&=& 2c_r+\\frac{r+1}{r}-\\frac{5c_r}{2} \\\\\n&=& 2c_r+\\frac{r^2-2r-8}{r(6r-8)} \\\\\n&\\ge& 2c_r;\n\\end{eqnarray*}\nsimilarly, since $c_r\\le\\frac{1}{2}$,\n\\begin{eqnarray*}\n\\frac{1}{r}\\sum_{i=1}^r \\frac{d(V_i)}{m} &>& \\frac{r+1}{r}-\\frac{c_r}{2} \\\\\n&>& \\frac{3}{4} \\\\\n&\\ge& \\frac{2}{3}+\\frac{c_r}{6};\n\\end{eqnarray*}\nso, since $c_r>\\frac{1}{3}$ for every $r$, Lemma \\ref{rulast}, following, ensures that $j\\ge k+2l$.\n\\end{proof}\n\nWe now give the result needed to fill in the gap.\n\n\\begin{lemma}\\label{rulast}\nLet $\\frac{1}{3}\\le c\\le \\frac{1}{2}$. If we have a finite, non-empty collection of numbers in $[0,1]$, whose \narithmetic mean $a$ is at least $\\max\\left(2c,\\frac{2}{3}+\\frac{c}{6}\\right)$, of which $j$ are at least $2c$, \n$k$ are at least $\\frac{c}{2}$ but less than $c$, $l$ are less than $\\frac{c}{2}$, and the rest are at least \n$c$ but less than $2c$, then $j\\ge k+2l$. \n\\end{lemma}\n\\begin{proof}\nWe proceed by induction on $k+l$; if $k=l=0$ then there is nothing to prove. If $k\\ge 1$ then one of our \nnumbers is less than the mean, so another must exceed the mean; since the mean of those \ntwo numbers is at most $\\frac{1+c}{2}\\le 2c \\le a$, either there are no other numbers (in which case we are \ndone) or the mean of the remaining numbers is at least $a$ and we are done by induction.\nIf $l\\ge 1$ then one number is less than $\\frac{c}{2}$; if also $j\\le 1$ then the total is less than \n$1+\\frac{c}{2}+(n-2)2c<2cn$, so there must be two numbers which are at least $2c$. Since the mean of these \nthree numbers is at most $\\frac{2}{3}+\\frac{c}{6}\\le a$, either there are no other numbers (in which case we \nare done) or the mean of the remaining numbers is at least $a$ and again we are done by induction.\n\\end{proof}\nThe value of $c_r$ which we use is tight only in the second case of Theorem \\ref{rumain} (for $r>4$); if we were to use instead some $c'_r > c_r$ it would be feasible for two of our parts to meet just under $c'_r$ edges while the remaining parts each met just under $2c'_r$ edges, giving us no immediate way to use Lemmas \\ref{rutwothree} and \\ref{rugoodbad}. To improve further on these bounds using a similar method, then, we might seek to prove analogues of Lemmas \\ref{rutwothree} and \\ref{rugoodbad} for parts which meet more than $\\beta cm$ for some $1<\\beta<2$. To do this, however, we would need a way to impose some stronger assumption on the degrees of vertices than simply $\\Delta(G)<cm$.\n\n", "meta": {"timestamp": "2017-01-23T02:07:45", "yymm": "1701", "arxiv_id": "1701.05855", "language": "en", "url": "https://arxiv.org/abs/1701.05855", "abstract": "The vertices of any graph with $m$ edges may be partitioned into two parts so that each part meets at least $\\frac{2m}{3}$ edges. Bollobás and Thomason conjectured that the vertices of any $r$-uniform hypergraph with $m$ edges may likewise be partitioned into $r$ classes such that each part meets at least $\\frac{r}{2r-1}m$ edges. In this paper we prove the weaker statement that, for each $r\\ge 4$, a partition into $r$ classes may be found in which each class meets at least $\\frac{r}{3r-4}m$ edges, a substantial improvement on previous bounds.", "subjects": "Combinatorics (math.CO)", "title": "Judicious partitions of uniform hypergraphs", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.987946220128863, "lm_q2_score": 0.8577680995361899, "lm_q1q2_score": 0.8474287516838972}}
{"url": "https://arxiv.org/abs/2204.05074", "title": "Supercritical Site Percolation on the Hypercube: Small Components are Small", "abstract": "We consider supercritical site percolation on the $d$-dimensional hypercube $Q^d$. We show that typically all components in the percolated hypercube, besides the giant, are of size $O(d)$. This resolves a conjecture of Bollobás, Kohayakawa, and Łuczak from 1994.", "text": "\\section{Introduction}\nThe $d$-dimensional hypercube $Q^d$ is the graph with the vertex set $V(Q^d)=\\{0,1\\}^d$, where two vertices are adjacent if they differ in exactly one coordinate. Throughout this paper, we denote by $n=2^d$ the order of the hypercube.\n\nIn bond percolation on $G$, one considers the subgraph $G_p$ obtained by including every edge independently with probability $p$. Erd\\H{o}s and Spencer \\cite{ES} conjectured that, similar to the classical $G(n,p)$ model, there is a phase transition at $p=\\frac{1}{d}$ in $Q^d_p$: for $\\epsilon>0$, when $p=\\frac{1-\\epsilon}{d}$, \\textbf{whp} all components have size $O(d)$, and when $p=\\frac{1+\\epsilon}{d}$, there exists \\textbf{whp} a unique giant component in $Q^d_p$ whose order is linear in $n$. Their conjecture was proved by Ajtai, Koml\\'os, and Szemer\\'edi \\cite{AKS}. Bollob\\'as, Kohayakawa, and Łuczak \\cite{BKL2} improved upon that result, extending it to a wider range of $p$, allowing $\\epsilon=o(1)$, and giving an asymptotic value of the order of the giant component of $Q^d_p$. Furthermore, they proved that the second largest component in $Q^d_p$ is typically of order $O(d)$.\n\nIn site percolation on $G$, one considers the induced subgraph $G[R]$, where $R$ is a random subset of vertices formed by including each vertex independently with probability $p$. In the setting of site percolation on the hypercube, Bollob\\'as, Kohayakawa, and Łuczak \\cite{BKL} proved the following:\n\n\n\\begin{thm}[Theorems 8, 9 of \\cite{BKL}]\\label{BKL}\n\\textit{Let $\\epsilon>0$ be a small enough constant, and let $p=\\frac{1+\\epsilon}{d}$. Let $R$ be a random subset formed by including each vertex of $Q^d$ independently with probability $p$. Then, \\textbf{whp}, there is a unique giant component of $Q^d[R]$, whose asymptotic size is $\\frac{\\left(2\\epsilon-O(\\epsilon^2)\\right)n}{d}$. Furthermore, \\textbf{whp} the size of the other components is at most $d^{10}$.}\n\\end{thm}\n\nMotivated by the results in bond percolation on the hypercube, in the same paper \\cite{BKL}, and also in their paper on bond percolation on the hypercube \\cite{BKL2}, Bollobás, Kohayakawa, and Łuczak conjectured that \\textbf{whp} all the components of $Q^d[R]$ besides the giant component are of asymptotic size at most $\\gamma d$, where $\\gamma$ is a constant depending only on $\\epsilon$ (see Conjecture 11 of \\cite{BKL}). We prove this conjecture:\n\\begin{theorem}\\label{main-theorem}\nLet $\\epsilon>0$ be a small enough constant, and let $p=\\frac{1+\\epsilon}{d}$. Let $R$ be a random subset formed by including each vertex of $Q^d$ independently with probability $p$. Then, \\textbf{whp}, all the components of $Q^d[R]$ besides the giant component are of size $O_{\\epsilon}\\left(d\\right)$.\n\\end{theorem}\n\nWe note that in our proof we obtain an inverse polynomial dependency on $\\epsilon$, of order $\\frac{1}{\\epsilon^5}$, in the hidden constant in $O_{\\epsilon}(d)$.\n\nThe text is structured as follows. In Section $2$, we show that big components are 'everywhere dense' (see the exact definition in Lemma \\ref{dense-giant-2}). In Section $3$, we exhibit several structures that \\textbf{whp} do not appear in the site-percolated hypercube. In Section $4$, we prove Theorem \\ref{main-theorem}.\n\nOur notation is fairly standard. Throughout the rest of the text, unless specifically stated otherwise, we set $\\epsilon$ to be a small enough constant, $p=\\frac{1+\\epsilon}{d}$, and $R$ to be a random subset of $V(Q^d)$ formed by including each vertex independently with probability $p$. Furthermore, we denote by $L_1$ the largest connected component of $Q^d[R]$. We denote by $N_G(S)$ the external neighbourhood of a set $S$ in the graph $G$. When $T$ is a subset of $V(G)$, we write $N_T(S)=N_G(S)\\cap T$. We omit rounding signs for the sake of clarity of presentation.\n\n\n\\section{Big components are everywhere dense}\nWe begin with the following claim, which holds for any $d$-regular graph:\n\\begin{claim} \\label{DFS-lemma}\n\\textit{Let $G=(V,E)$ be a $d$-regular graph on $n$ vertices. Let $\\epsilon$ be a small enough constant, and let $p=\\frac{1+\\epsilon}{d}$. Form $R\\subseteq V$ by including each vertex of $V$ independently with probability $p$. Then, \\textbf{whp}, any connected component $S$ of $G[R]$ with $|S|=k>300\\ln n$ has $|N_G(S)|\\ge \\frac{9kd}{10}$.}\n\\end{claim}\n\\begin{proof}\nWe run the Depth First Search (DFS) algorithm with $n$ random bits $X_i$, discovering the connected components while generating $G[R]$ (see \\cite{site-percolation}). If there is a connected component $S$ of size $k$, then there is an interval of random bits whose length is $k+|N_G(S)|$ in the DFS where we receive $k$ positive answers. Thus, the probability of having a component violating the claim, whose discovery by the DFS starts with a given bit $X_i$ is at most:\n\\begin{align*}\n    P\\left[Bin\\left(\\frac{9kd}{10}+k,\\frac{1+\\epsilon}{d}\\right)\\ge k\\right]\n\\end{align*}\nHence, by a Chernoff-type bound (see Appendix A in \\cite{probablistic method}), the probability of having such a component is at most\n\\begin{align*}\n    \\exp\\left(-\\frac{k}{100}\\right)\\le o\\left(\\frac{1}{n^2}\\right),\n\\end{align*}\nwhere the inequality follows from our bound on $k$. We complete the proof with a union bound on the $<n$ possible values of $k$ and $< n$ possible starting points of the interval in the DFS.\n\\end{proof}\n\n\nWe further require the following simple claim:\n\\begin{claim} \\label{seperating}\nLet $S\\subseteq V(Q^d)$ such that $|S|=k\\le d$. Then, there exist pairwise disjoint subcubes, each of dimension at least $d-k+1$, such that every $v\\in S$ is in exactly one of these subcubes. \n\\end{claim}\n\\begin{proof}\nWe prove by induction on $k$. The case $k=1$ is trivial.\n\nAssume that the statement holds for all $k'<k$. Choose any two vertices of $S$. There is at least one coordinate on which they do not agree. Consider the two pairwise disjoint (and complementary) subcubes of dimension $d-1$: one where we fix this coordinate to be $0$ and let the other coordinates vary, and the other where we fix this coordinate to be $1$ and let the other coordinates vary. Clearly, each of the two vertices is in exactly one of these subcubes. Since these subcubes are disjoint, no vertex from the other $k-2$ vertices is in both of them. We can then apply the induction hypothesis on each of these subcubes, giving rise to pairwise disjoint subcubes of dimension at least $d-1-(k-2)=d-k+1$, each having exactly one of the vertices of $S$.\n\\end{proof}\n\n\nWe conclude this section with a lemma, bounding the probability that a fixed small set of vertices has no vertex at Hamming distance $1$ from many vertices in large components or in their neighbourhoods. The proof here borrows several ideas used in \\cite{AKS, BKL, EKK}.\n\\begin{lemma}\\label{dense-giant-2}\nFix $S\\subseteq V(Q^d)$ such that $|S|=k\\le \\frac{\\epsilon d}{10}$. Then, the probability that every $v\\in S$ has less than $\\frac{\\epsilon^2d}{40}$ neighbours (in $Q^d$) in components of $Q^d[R]$ whose size is at least $n^{1-\\epsilon}$ or in their neighbourhoods in $Q^d$ is at most $\\exp\\left(-\\frac{\\epsilon^2dk}{40}\\right)$.\n\\end{lemma}\n\\begin{proof}\nLet $S=\\{v_1,\\cdots, v_k\\}$. By Claim \\ref{seperating}, we can consider pairwise disjoint subcubes of dimension at least $\\left(1-\\frac{\\epsilon }{10}\\right)d$, each containing exactly one of the vertices of $S$. We denote these subcubes by $Q_1,\\cdots, Q_k$, where $v_i\\in V(Q_i)$. \n\nFor each $i$, consider $Q_i$, and assume, without loss of generality, that $v_i$ is the origin of $Q_i$. We then create $\\frac{\\epsilon d}{10}$ pairwise disjoint subcubes of $Q_i$ of dimension at least $\\left(1-\\frac{\\epsilon}{5}\\right)d$, each at distance $1$ from $v_i$, by fixing one of the first $\\frac{\\epsilon d}{10}$ coordinates of $Q_i$ to be $1$, the rest of the first $\\frac{\\epsilon d}{10}$ coordinates to be $0$, and letting the other coordinates vary. Denote these subcubes by $Q_i(1), \\cdots, Q_i\\left(\\frac{\\epsilon d}{10}\\right)$, and the vertex at distance $1$ from $v_i$ in $Q_i(j)$ by $v_i(j)$. We denote by $n'$ the order of each $Q_i(j)$, and note that $n'\\ge2^{\\left(1-\\frac{\\epsilon}{5}\\right)d}=n^{1-\\frac{\\epsilon}{5}}$.\n\nFix $i$. Observe that $p$ is still supercritical at every $Q_i(j)$ since $(1+\\epsilon)\\left(1-\\frac{\\epsilon}{5}\\right)\\ge 1+\\frac{3\\epsilon}{5}$. Denote by $L_1(i,j)$ the largest component of $Q_i(j)[R]$. Then, by Theorem \\ref{BKL}, \\textbf{whp} $\\big|L_1(i,j)\\big|\\ge \\frac{7\\epsilon n'}{6d}.$ Furthermore, by Claim \\ref{DFS-lemma}, \\textbf{whp}\n$\\bigg|N_{Q_i(j)}\\left(L_1(i,j)\\right)\\bigg|\\ge \\frac{21\\epsilon n'}{20}.$ Thus, setting $\\mathcal{A}_{(i,j)}$ to be the event that $\\bigg|L_1(i,j)\\cup N_{Q_i(j)}\\left(L_1(i,j)\\right)\\bigg|\\ge\\frac{21\\epsilon n'}{20}$, we have that $P\\left(\\mathcal{A}_{(i,j)}\\right)=1-o(1)$. Now, let $\\mathcal{B}_{(i,j)}$ be the event that $v_i(j)\\in L_1(i,j)\\cup N_{Q_i(j)}\\left(L_1(i,j)\\right)$. Since the hypercube is transitive, we have that $P\\left(\\mathcal{B}_{(i,j)}|\\mathcal{A}_{(i,j)}\\right)\\ge \\frac{21\\epsilon n'/20}{n'}=\\frac{21\\epsilon}{20}$. Therefore, $P\\left(\\mathcal{B}_{(i,j)}\\cap \\mathcal{A}_{(i,j)}\\right)\\ge \\left(1-o(1)\\right)\\frac{21 \\epsilon}{20}\\ge \\epsilon$. Since the subcubes $Q_i(j)$ are pairwise disjoint, the events $B_{(i,j)}\\cap \\mathcal{A}_{(i,j)}$ are independent for different $j$. Thus, by a typical Chernoff-type bound, with probability at least $1-\\exp\\left(-\\frac{\\epsilon^2d}{40}\\right)$, at least $\\frac{\\epsilon^2d}{40}$ of the $v_i(j)$ are in $L_1(i,j)\\cup N_{Q_i(j)}\\left(L_1(i,j)\\right)$ with $\\bigg|L_1(i,j)\\cup N_{Q_i(j)}\\left(L_1(i,j)\\right)\\bigg|\\ge\\frac{21\\epsilon n'}{20}$. Thus, with the same probability, $v_i$ is at distance $1$ from at least $\\frac{\\epsilon^2d}{40}$ vertices in components whose size is at least $\\frac{\\epsilon n'}{d}\\ge \\frac{\\epsilon n^{1-\\frac{\\epsilon}{5}}}{d}\\ge n^{1-\\epsilon}$ or in their neighbourhoods.\n\nSince $v_i$'s are in pairwise disjoint subcubes, these events are also independent for each $v_i$. Hence, the probability that none of the $v_i$ are at distance $1$ from at least $\\frac{\\epsilon^2d}{40}$ vertices in components whose size is at least $n^{1-\\epsilon}$ or in their neighbourhoods is at most $\\exp\\left(-\\frac{\\epsilon^2dk}{40}\\right)$.\n\\end{proof}\n\n\\section{Unlikely structures in the percolated hypercube}\n\nDenote by $N^2_G(v)$ the set of vertices in $G$ at distance exactly $2$ from $v$. The following lemma shows that, typically, there are no large sections of a sphere of radius $2$ in $Q^d[R]$.\n\\begin{lemma}\\label{no-sphere}\n\\textbf{Whp}, there is no $v\\in Q^d$ such that $\\big|N^2_{Q^d}(v)\\cap R\\big|\\ge 2d$.\n\\end{lemma}\n\\begin{proof}\nWe have $n$ ways to choose $v$, and $\\binom{\\binom{d}{2}}{2d}$ ways to choose a subset of $2d$ vertices from $N^2_{Q^d}(v)$. We include them in $R$ with probability at most $p^{2d}$. Hence, by the union bound, the probability of violating the lemma is at most:\n\\begin{align*}\n    2^d\\binom{\\binom{d}{2}}{2d}\\left(\\frac{1+\\epsilon}{d}\\right)^{2d}\\le 2^d\\left(\\frac{ed}{4}\\cdot \\frac{1+\\epsilon}{d}\\right)^{2d}\\le \\left(\\frac{14}{15}\\right)^{d}=o(1).\n\\end{align*}\n\\end{proof}\n\nFrom Lemma \\ref{no-sphere}, we are able to derive the following lemma:\n\\begin{lemma} \\label{big-neighbourhood}\n\\textbf{Whp} there are no $S\\subseteq R$ and $W\\subseteq V(Q^d)$ disjoint from $S$ such that $|W|\\le\\frac{\\epsilon^4d|S|}{9\\cdot 200^2}$ and every $v\\in S$ has $d_{Q^d}(v, W)\\ge \\frac{\\epsilon^2d}{200}$.\n\\end{lemma}\n\\begin{proof}\nAssume the contrary. By our assumption on $S$, there are at least $\\frac{\\epsilon^2 d|S|}{200}$ edges between $S$ and $W$. Thus, the average degree from $W$ to $S$ is at least $\\frac{\\epsilon^2 d|S|}{200|W|}$. Now, let us count the number of cherries with the vertex of degree $2$ in $W$. By Jensen's inequality, we have at least $\\binom{\\epsilon^2 d|S|/200|W|}{2}|W|\\ge \\frac{\\epsilon^4d^2|S|^2}{4\\cdot 200^2|W|}\\ge \\frac{9d|S|}{4}$ such cherries, where we used our assumption on $|W|$. Thus, by the pigeonhole principle, there is a vertex $v\\in S$ that is in $\\frac{2}{|S|}\\cdot\\frac{9d|S|}{4}=\\frac{9d}{2}$ cherries. Since every pair of vertices in $S$ is connected by at most two paths of length $2$ in $Q^d$, we obtain that $v$ is at Hamming distance $2$ from at least $\\frac{\\frac{9d}{2}}{2}=\\frac{9d}{4}>2d$ vertices in $S\\subseteq R$. On the other hand, by Lemma \\ref{no-sphere}, \\textbf{whp} there is no $v\\in Q^d$ such that $|N_{Q^d}^2(v)\\cap R|\\ge 2d$, completing the proof.\n\\end{proof}\n\n\nThe next lemma bounds the number of subtrees of a given order in a $d$-regular graph.\n\\begin{lemma} \\label{trees}\n\\textit{Let $t_k(G)$ denote the number of $k$-vertex trees contained in a $d$-regular graph $G$ on $n$ vertices. Then,\n$$t_k(G)\\le n(ed)^{k-1}.$$}\n\\end{lemma}\nThis follows directly from Lemma 2.1 of \\cite{trees}.\n\nWe are now ready to state and prove the final lemma of this section:\n\\begin{lemma}\\label{no-squid}\nLet $C>0$ be a constant. Then, \\textbf{whp}, there is no $S\\subseteq V(Q^d)$, such that $|S|\\le Cd$, $S$ is connected in $Q^d$, and at least $\\frac{\\epsilon d}{10}$ vertices $v\\in S$ have that $\\big|N_{L_1\\cup N_{Q^d}(L_1)}(v)\\big|<\\frac{\\epsilon^2d}{40}$.\n\\end{lemma}\n\\begin{proof}\nBy Theorem \\ref{BKL}, \\textbf{whp} there is a unique giant component whose size is larger than $d^{10}$, which we denote by $L_1$. Thus, it suffices to show that \\textbf{whp} there is no such $S$, where $\\frac{\\epsilon d}{10}$ of its vertices have less than $\\frac{\\epsilon^2d}{40}$ vertices at Hamming distance $1$ from components of size at least $n^{1-\\epsilon}$ or their neighbourhoods, since \\textbf{whp} these components and their neighbourhoods are in fact $L_1\\cup N_{Q^d}(L_1)$. \n\nSince $Q^d$ is connected, we can consider connected sets of size exactly $Cd$. By Lemma \\ref{trees}, we have $n(ed)^{Cd}$ ways to choose $S$. We have $\\binom{Cd}{\\frac{\\epsilon d}{10}}$ ways to choose the vertices in $S$ which do not have at least $\\frac{\\epsilon^2d}{40}$ vertices at Hamming distance $1$ from components of size at least $n^{1-\\epsilon}$ or their neighbourhoods. By Lemma \\ref{dense-giant-2}, the probability that no vertex in a given set of $\\frac{\\epsilon d}{10}$ vertices in $S$ has at least $\\frac{\\epsilon^2d}{40}$ vertices at Hamming distance $1$ from components of size at least $n^{1-\\epsilon}$ or their neighbourhoods is at most $\\exp\\left(-\\frac{\\epsilon^4d^2}{400}\\right)$. Hence, the probability of the event violating the statement of the lemma is at most:\n\\begin{align*}\n    n(ed)^{Cd}\\binom{Cd}{\\frac{\\epsilon d}{10}}\\exp\\left(-\\frac{\\epsilon^4d^2}{400}\\right)\\le n\\left((ed)^C\\left(\\frac{10eC}{\\epsilon}\\right)^{\\frac{\\epsilon}{10}}\\exp\\left(-\\frac{\\epsilon^4d}{400}\\right)\\right)^d=o(1).\n\\end{align*} \n\\end{proof}\n\n\\section{Proof of Theorem \\ref{main-theorem}}\nLet $p_1=\\frac{1+\\epsilon/2}{d}$. Form $R_1$ by including each vertex $v\\in V(Q^d)$ independently with probability $p_1$. By Theorem \\ref{BKL}, \\textbf{whp} there is a unique giant component, denote it by $L_1'$. We can thus split the vertices of the hypercube into the following three disjoint sets: $T=L_1'\\cup N_{Q^d}(L_1')$, $M$ is the set of vertices outside $T$ with at least $\\frac{\\epsilon^2d}{200}$ neighbours in $T$, and $S=V\\left(Q^d\\right)\\setminus\\left(T\\cup M\\right)$.\n\nLet $p_2=\\frac{\\epsilon}{2d-2-\\epsilon}$. Form $R_2$ by including each vertex $v\\in V(Q^d)$ independently with probability $p_2$. Note that since $1-p=(1-p_1)(1-p_2)$, the random set $R$ has the same distribution as $R_1\\cup R_2$. Thus, with a slight abuse of notation, we write $R=R_1\\cup R_2$. We begin by performing the second exposure (i.e. by generating $R_2$) on $S\\cup M$, and only afterwards on $T$. Note that once we show that a connected set has a neighbour in $T\\cap R_2$, it means that it merges into $L_1'$ (whereas by Theorem \\ref{BKL}, \\textbf{whp} $L_1'\\subseteq L_1$).\n\n\nLet $C_1$ be a positive constant, possibly depending on $\\epsilon$. We first show that \\textbf{whp} if there is a connected component $B$ in $(S\\cup M)\\cap R$ such that $\\big|B\\cap M\\big|\\ge C_1d$, it merges with $L_1'$ after the second exposure on $T$. By construction, every $v\\in M$ has at least $\\frac{\\epsilon^2d}{200}$ neighbours in $T$, and $T\\cap M=\\emptyset$. Thus, by Lemma \\ref{big-neighbourhood}, we have that \\textbf{whp} $\\big|N_T(B\\cap M)\\big|\\ge \\frac{C_1\\epsilon^4d^2}{9\\cdot 200^2}$. Hence,\n\\begin{align*}\n    P\\left[\\big|N_T(B)\\cap R_2\\big|=0\\right]&\\le \\left(1-\\frac{\\epsilon}{2d}\\right)^{\\frac{C_1\\epsilon^4d^2}{9\\cdot200^2}}\\le \\exp\\left(-\\frac{C_1\\epsilon^5d}{18\\cdot200^2}\\right)=o(1/n),\n\\end{align*}\nby choosing $C_1\\ge\\frac{18\\cdot200^2}{\\epsilon^5}$. We conclude with the union bound over the $<n$ connected components in $(S\\cup M)\\cap R$.\n\nThus, we are left to deal with connected components $B\\subseteq(S\\cup M)\\cap R$ of size at least $2C_1d$ with less than $C_1d$ vertices in $B\\cap M$. Taking a connected subset $B'\\subseteq B$ of size exactly $Cd:=2C_1d$, we have that $|B'\\cap S|\\ge |B'|-|B'\\cap M|\\ge C_1d\\ge \\frac{\\epsilon d}{20}$. But by Lemma \\ref{no-squid}, \\textbf{whp} there is no connected set of size $Cd$ in $Q^d$ with at least $\\frac{\\frac{\\epsilon}{2} d}{10}=\\frac{\\epsilon d}{20}$ vertices which do not have at least $\\frac{\\left(\\frac{\\epsilon}{2}\\right)^2d}{40}\\ge \\frac{\\epsilon^2d}{200}$ vertices at Hamming distance $1$ from $L_1'\\cup N_{Q^d}(L_1')$. Recalling that $S, M,T$ and $L_1'$ were formed according to the supercritical percolation with $p_1=\\frac{1+\\epsilon/2}{d}$, this means that \\textbf{whp} there is no such connected set $B'$. All in all, \\textbf{whp}, there is no connected component $B$ in $Q^d[R]$ of size at least $Cd$ outside of $L_1$. \\qedsymbol\n\n\\paragraph{Acknowledgements.} The first author wishes to thank Arnon Chor and Dor Elboim for fruitful discussions.\n\n\n", "meta": {"timestamp": "2022-04-12T02:40:58", "yymm": "2204", "arxiv_id": "2204.05074", "language": "en", "url": "https://arxiv.org/abs/2204.05074", "abstract": "We consider supercritical site percolation on the $d$-dimensional hypercube $Q^d$. We show that typically all components in the percolated hypercube, besides the giant, are of size $O(d)$. This resolves a conjecture of Bollobás, Kohayakawa, and Łuczak from 1994.", "subjects": "Probability (math.PR); Combinatorics (math.CO)", "title": "Supercritical Site Percolation on the Hypercube: Small Components are Small", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9835969684454966, "lm_q2_score": 0.8615382040983515, "lm_q1q2_score": 0.847406365751116}}
{"url": "https://arxiv.org/abs/1109.6277", "title": "Domination Value in Graphs", "abstract": "A set $D \\subseteq V(G)$ is a \\emph{dominating set} of $G$ if every vertex not in $D$ is adjacent to at least one vertex in $D$. A dominating set of $G$ of minimum cardinality is called a $\\gamma(G)$-set. For each vertex $v \\in V(G)$, we define the \\emph{domination value} of $v$ to be the number of $\\gamma(G)$-sets to which $v$ belongs. In this paper, we study some basic properties of the domination value function, thus initiating \\emph{a local study of domination} in graphs. Further, we characterize domination value for the Petersen graph, complete $n$-partite graphs, cycles, and paths.", "text": "\\section{Introduction}\n\nLet $G = (V(G),E(G))$ be a simple, undirected, and nontrivial\ngraph with order $|V(G)|$ and size $|E(G)|$. For $S \\subseteq\nV(G)$, we denote by $\\langle S \\rangle$ the subgraph of $G$ induced by $S$.\nThe \\emph{degree of a vertex $v$} in $G$, denoted by $\\deg_G(v)$,\nis the number of edges that are incident to $v$ in $G$; an \\emph{end-vertex} is a vertex of degree one, \nand a \\emph{support vertex} is a vertex that is adjacent to an end-vertex. We denote by\n$\\Delta(G)$ \\emph{the maximum degree} of a graph $G$. For a vertex\n$v \\in V(G)$, the \\emph{open neighborhood} $N(v)$ of $v$ is the\nset of all vertices adjacent to $v$ in $G$, and the \\emph{closed\nneighborhood}  $N[v]$ of $v$ is the set $N(v) \\cup \\{v\\}$. A set $D\n\\subseteq V(G)$ is a \\emph{dominating set} (DS) of $G$ if for each\n$v \\not\\in D$ there exists a $u \\in D$ such that $uv\\in E(G)$. The\n\\emph{domination number} of $G$, denoted by $\\gamma(G)$, is the\nminimum cardinality of a DS in $G$; a DS of $G$ of minimum\ncardinality is called a $\\gamma(G)$-set. For earlier discussions\non domination in graphs, see \\cite{B1, B2, EJ, Jaegar, Ore}. For a\nsurvey of domination in graphs, refer to \\cite{Dom1, Dom2}. We generally\nfollow \\cite{CZ} for notation and graph theory terminology.\nThroughout the paper, we denote by $P_n$, $C_n$, and $K_n$ the path,\nthe cycle, and the complete graph on $n$ vertices, respectively.\\\\\n\nIn \\cite{Slater}, Slater introduced the notion of the number of dominating sets of $G$, which he denoted by HED$(G)$ in honor of Steve Hedetniemi; further, he also \nused \\#$\\gamma(G)$ to denote the number of $\\gamma(G)$-sets. In this paper, we will use $\\tau(G)$ to denote the total number of $\\gamma(G)$-sets, and\nby $DM(G)$ the collection of all $\\gamma(G)$-sets. For each vertex\n$v \\in V(G)$, we define the \\emph{domination value} of $v$,\ndenoted by $DV_G(v)$, to be the number of $\\gamma(G)$-sets to\nwhich $v$ belongs; we often drop $G$ when ambiguity is not a\nconcern. See \\cite{CK} for a discussion on \\emph{total domination value} in graphs. For a further work on domination value in graphs, see \\cite{Yi}. In this paper, we study some basic\nproperties of the domination value function, thus initiating a\n\\emph{local study} of domination in graphs. When a real-world\nsituation can be modeled by a graph, the locations (vertices)\nwith high domination values are of interest. One can use\ndomination value in selecting locations for fire departments or\nconvenience stores, for example. Though numerous papers on\ndomination have been published, no prior systematic local study of\ndomination is known. However, in \\cite{pruned tree}, Mynhardt\ncharacterized the vertices in a tree $T$ whose domination value is $0$ or $\\tau(T)$. \nIt should be noted that finding domination value of any given vertex in \na given graph $G$ can be an extremely difficulty task, given the difficulty attendant to finding $\\tau(G)$ or just $\\gamma(G)$.\\\\\n\n\n\n\\section{Basic properties of domination value: upper and lower bounds}\n\nIn this section, we consider the lower and upper bounds of the\ndomination value function for a fixed vertex $v_0$ and for $v \\in\nN[v_0]$. Clearly, $0 \\le DV_G(v) \\le \\tau(G)$ for any graph $G$\nand for any vertex $v \\in V(G)$. We will say the bound is sharp if\nequality is obtained for a graph of some order in an inequality.\nWe first make the following observations.\\\\\n\n\\begin{observation}\\label{observation}\n$\\displaystyle \\sum_{v \\in V(G)} DV_G(v) = \\tau(G) \\cdot\n\\gamma(G)$\n\\end{observation}\n\n\\begin{observation}\\label{observation1}\nIf there is an isomorphism of graphs carrying a vertex $v$ in $G$\nto a vertex $v'$ in $G'$, then $DV_G(v)=DV_{G'}(v')$.\n\\end{observation}\n\nExamples of graphs that admit automorphisms are cycles, paths, and\nthe Petersen graph. The Pertersen graph, which is often used as a\ncounter-example for conjectures, is vertex-transitive (p.27,\n\\cite{Petersen}). Let $\\mathcal{P}$ denote the Petersen graph with\nlabeling as in Figure \\ref{P}.\n\n\\begin{figure}[htbp]\n\\begin{center}\n\\scalebox{0.5}{\\input{P.pstex_t}} \\caption{The Petersen\ngraph}\\label{P}\n\\end{center}\n\\end{figure}\n\nIt's easy to check that $\\gamma(\\mathcal{P})=3$. We will show that\n$DV(v)=3$ for each $v \\in V(\\mathcal{P})$. Since $\\mathcal{P}$ is\nvertex-transitive, it suffices to compute $DV_{\\mathcal{P}}(1)$.\nFor the $\\gamma(\\mathcal{P})$-set $\\Gamma$ containing the vertex\n$1$, one can easily check that no vertex in $N(1)$ belongs to\n$\\Gamma$. Further, notice that no three vertices from\n$\\{1,2,3,4,5\\}$ form a $\\gamma(\\mathcal{P})$-set. Keeping these\ntwo conditions in mind, one can readily verify that the\n$\\gamma(\\mathcal{P})$-sets containing the vertex $1$ are\n$\\{1,3,7\\}$, $\\{1,4,10\\}$, and $\\{1,8,9\\}$, and thus\n$DV(1)=3=DV(v)$ for each $v \\in V(\\mathcal{P})$.\n\n\n\\begin{observation}\\label{observation2}\nLet $G$ be the disjoint union of two graphs $G_1$ and $G_2$. Then\n$\\gamma(G)=\\gamma(G_1)+ \\gamma(G_2)$ and $\\tau(G)=\\tau(G_1) \\cdot\n\\tau(G_2)$. For $v \\in V(G_1)$, $DV_G(v)=DV_{G_1}(v) \\cdot\n\\tau(G_2)$.\n\\end{observation}\n\n\\begin{proposition}\\label{upperbound1}\nFor a fixed $v_0 \\in V(G)$, we have\n$$\\tau(G) \\le \\sum_{v \\in N[v_0]} DV_G(v) \\le \\tau(G) \\cdot \\gamma(G),$$ and both bounds are\nsharp.\n\\end{proposition}\n\n\\begin{proof}\nThe upper bound follows from Observation \\ref{observation}. For\nthe lower bound, note that every $\\gamma(G)$-set $\\Gamma$\nmust contain a vertex in $N[v_0]$: otherwise $\\Gamma$ fails to dominate $v_0$.\\\\\n\nFor sharpness of the lower bound, take $v_0$ to be an end-vertex\nof $P_{3k}$ for $k \\ge 1$ (see Theorem \\ref{theorem on paths} and\nCorollary \\ref{path on 3k}). For sharpness of the upper bound,\ntake as $v_0$ the central vertex of (A) in Figure \\ref{figure1}. \\hfill\n\\end{proof}\n\n\\begin{proposition} \\label{upperbound2}\nFor any $v_0 \\in V(G)$, $$\\sum_{v \\in N[v_0]} DV_G(v) \\le \\tau(G)\n\\cdot (1+ \\deg_G(v_0)),$$ and the bound is sharp.\n\\end{proposition}\n\n\\begin{proof}\nFor each $v\\in N[v_0]$, $DV_G(v)\\leq \\tau(G)$ and $|N[v_0]|=1+ \\deg_G(v_0)$. Thus,\\\\\n$$\\sum_{v\\in N[v_0]}DV(v) \\leq \\sum_{v\\in N[v_0]}\\tau(G)=\\tau(G)\\!\\!\\!\\sum_{v\\in N[v_0]}1=\\tau(G)(1+ \\deg_G(v_0)).$$\\\\\nThe upper bound is achieved for a graph of order $n$ for any\n$n\\geq 1$. Let $G_n$ be a graph on $n$ vertices containing an\nisolated vertex. To see the sharpness of the upper bound, take as $v_0$\none of the isolates vertices, then the upper bound follows by\nObservation \\ref{observation2} and $\\deg_G(v_0)=0$. \\hfill\n\\end{proof}\n\n\\begin{figure}[htbp]\n\\begin{center}\n\\scalebox{0.5}{\\input{DVupperbound.pstex_t}} \\caption{Examples of\nlocal domination values and their upper bounds}\\label{figure1}\n\\end{center}\n\\end{figure}\n\nWe will compare two examples, where each example attains the upper\nbound of Proposition \\ref{upperbound1} or Proposition\n\\ref{upperbound2}, but not both. Let $v_0$ be the central vertex\nof degree $3$, which is not a support vertex as in the graph (A)\nof Figure \\ref{figure1}. Then $\\sum_{v \\in N[v_0]} DV_G(v)=3$.\nNote that $\\tau (G) =1$, $\\gamma (G)=3$, and $\\deg_G(v_0)=3$.\nProposition \\ref{upperbound1} yields the upper bound $\\tau (G)\n\\cdot \\gamma (G)=1\\cdot 3=3$, which is sharp. But, the upper bound\nprovided by Proposition \\ref{upperbound2}\nis $\\tau (G) \\cdot (1+\\deg_G(v_0))=1 \\cdot (1+3)=4$, which is not sharp in this case.\\\\\n\nNow, let $v_0$ be an isolated vertex as labeled in the graph (B)\nof Figure \\ref{figure1}. Then $\\sum_{v \\in N[v_0]}DV_{G'}(v)=2$.\nNote that $\\tau (G')=2$, $\\gamma(G')=4$, and $\\deg_{G'}(v_0)=0$.\nProposition \\ref{upperbound2} yields the upper bound $\\tau (G')\n\\cdot (1+ \\deg_{G'}(v_0))=2 \\cdot (1+0)=2$, which is sharp. But,\nthe upper bound provided by Proposition \\ref{upperbound1} is\n$\\tau (G') \\cdot \\gamma (G')=2 \\cdot 4=8$, which is not sharp in this case.\\\\\n\n\\begin{proposition}\\label{subgraph-tau}\nLet $H$ be a subgraph of $G$ with $V(H)=V(G)$. If\n$\\gamma(H)=\\gamma(G)$, then $\\tau(H) \\le \\tau(G)$.\n\\end{proposition}\n\n\\begin{proof}\nBy the first assumption, every DS for $H$ is a DS for $G$. By\n$\\gamma(H)=\\gamma(G)$, it's guaranteed that every DS of minimum\ncardinality for $H$ is also a DS \\emph{of minimum cardinality for $G$}.\n\\end{proof}\n\nThe \\emph{complement} $\\overline{G}=(V(\\overline{G}),\nE(\\overline{G}))$ of a graph $G$ is the graph such that\n$V(\\overline{G})=V(G)$ and $uv \\in E(\\overline{G})$ if and only if\n$uv \\not\\in E(G)$. We recall the following\n\n\\begin{theorem}\nLet $G$ be any graph of order $n$. Then\n\\begin{itemize}\n\\item[(i)] (\\cite{Jaegar}, Jaegar and Payan)\n$\\gamma(G)+\\gamma(\\overline{G}) \\le n+1$; and\n\\item[(ii)](\\cite{B1}, p.304) $\\gamma(G) \\le n- \\Delta(G)$.\n\\end{itemize}\n\\end{theorem}\n\n\\begin{proposition}\nLet $G$ be a graph on $n=2m \\ge 4$ vertices. If $G$ or\n$\\overline{G}$ is $mK_2$, then\n$$DV_G(v)+DV_{\\overline{G}}(v)=n-1+2^{\\frac{n}{2}-1}.$$\n\\end{proposition}\n\n\\begin{proof}\nWithout loss of generality, assume $G=mK_2$ and label the vertices\nof $G$ by $1,2,\\ldots,2m$. Further assume that the vertex $2k-1$\nis adjacent to the vertex $2k$, where $1 \\le k \\le m$. Then\n$DV_G(1)=2^{m-1}$, which consists of the vertex 1 and one vertex\nfrom each path $K_2$. By Observation \\ref{observation1} and\nObservation \\ref{observation2}, $DV_G(v)=2^{m-1}=2^{\\frac{n}{2}-1}$ for any $v\\in V(G)$.\\\\\n\nNow, consider $\\overline{G}$ and the vertex labeled $1$ for ease\nof notation. Since $\\Delta(G)=n-2$, $\\gamma(\\overline{G})>1$.\nNoting that $\\{1, \\alpha\\}$ as $\\alpha$ ranges from 2 to $2m$\nenumerates all dominating sets of $\\overline{G}$ containing the\nvertex 1, we have $\\gamma(\\overline{G})=2$ and\n$DV_{\\overline{G}}(1)=n-1$. By Observation \\ref{observation1},\n$DV_{\\overline{G}}(v)=n-1$ holds for any $v \\in V(\\overline{G})$.\nTherefore, $DV_G(v)+DV_{\\overline{G}}(v)=n-1+2^{\\frac{n}{2}-1}$. \\hfill\n\\end{proof}\n\nNext we consider domination value of a graph $G$ when $\\Delta(G)$ is\ngiven.\n\n\\begin{observation} \\label{degree n-1}\nLet $G$ be a graph of order $n \\ge 2$ such that $\\Delta (G)=n-1$. Then\n$\\gamma(G)=1$ and $DV(v) \\le 1$ for any $v \\in V(G)$. Equality\nholds if and only if $\\deg_G(v)=n-1$.\n\\end{observation}\n\n\\begin{proposition}\\label{degree n-2}\nLet $G$ be a graph of order $n \\ge 3$ such that $\\Delta(G)=n-2$. Then\n$\\gamma(G)=2$ and $DV(v) \\leq n-1$ for any $v \\in V(G)$. Further,\nif $\\deg(v)=n-2$, then $DV(v)=|N[w]|$ where $vw \\notin E(G)$.\n\\end{proposition}\n\n\\begin{proof}\nLet $\\deg_G(v)=\\Delta(G)=n-2$, then $\\gamma(G)>1$ and there's only\none vertex $w$ such that $vw \\notin E(G)$. Clearly, $\\{v, w\\}$ is\na $\\gamma(G)$-set; so $\\gamma(G)=2$. Noticing that $v$ dominates\n$N[v]$, we see that the number of $\\gamma(G)$-sets containing $v$\nis $|N[w]|$; i.e., $DV(v)=|N[w]| \\le n-2$. \\hfill\n\\end{proof}\n\n\\begin{theorem}\\label{theorem on Delta(G)=n-3}\nLet $G$ be a graph of order $n \\ge 4$ and $\\Delta (G)=n-3$. Fix a\nvertex $v$ with $\\deg_G(v)= \\Delta(G)$.\n\\begin{itemize}\n\\item [(i)] If $G$ is disconnected, then $\\gamma(G)=2$ with\n$DV(v)=2$ or $\\gamma(G)=3$ with $DV(v)\\le n-3$. \\item [(ii)] If\n$G$ is connected, then $\\gamma(G)=2$ with $DV(v) \\le n-2$ or\n$\\gamma(G)=3$ with $DV(v) \\le (\\frac{n-1}{2})^2$.\n\\end{itemize}\n\\end{theorem}\n\n\\begin{proof}\nSince $\\deg_G(v)=\\Delta(G)=n-3$, there are two vertices, say\n$\\alpha$ and $\\beta$, such that $v \\alpha, v \\beta \\not\\in E(G)$.\nWe consider four cases.\n\n\\begin{figure}[htbp]\n\\begin{center}\n\\scalebox{0.5}{\\input{max_n_3D.pstex_t}} \\caption{Cases 1, 2, and\n3 when $\\Delta(G)=n-3$}\\label{max1}\n\\end{center}\n\\end{figure}\n\n\\emph{Case 1. Neither $\\alpha$ nor $\\beta$ is adjacent to any\nvertex in $N[v]$:} Let $G^{\\prime}= \\langle V(G) - \\{\\alpha,\n\\beta\\} \\rangle$. Then $\\deg_{G'}(v)=n-3$ with $|V(G^{\\prime})|=n-2$.\nBy Observation~\\ref{degree n-1}, $\\gamma(G')=1$ and\n$DV_{G'}(v)=1$. First suppose $\\alpha$ and $\\beta$ are isolated\nvertices in $G$. (Consider (A) of Figure \\ref{max1} with the edge\n$\\alpha \\beta$ being removed.) Observation~\\ref{observation2},\ntogether with $\\gamma(\\langle\\{\\alpha,\\beta\\}\\rangle)=2$ and\n$\\tau(\\langle\\{\\alpha,\\beta\\}\\rangle)=1$, yields $\\gamma(G)=3$ and\n$DV_G(v)=1$. Next assume that $G$ has no isolated vertex, then $\\alpha\n\\beta \\in E(G)$ (see (A) of Figure \\ref{max1}). Observation\n\\ref{observation2}, together with $\\gamma(\\langle\\{\\alpha,\\beta \\}\\rangle)=1$ and\n$\\tau(\\langle\\{\\alpha,\\beta \\}\\rangle)=2$, yields $\\gamma(G)=2$ and $DV_G(v)=2$.\\\\\n\n\n\\emph{Case 2. Exactly one of $\\alpha$ and $\\beta$ is adjacent to a\nvertex in $N(v)$:} Without loss of generality, assume that\n$\\alpha$ is adjacent to a vertex in $N(v)$. First suppose that $G$\nis not connected. Then $\\alpha \\beta \\not\\in E(G)$. (Consider (B)\nof Figure \\ref{max1} with the edge $\\alpha \\beta$ being removed.)\nLet $G'=\\langle V(G)-\\{\\beta\\} \\rangle$. Then $\\deg_{G'}(v)=n-3$ with\n$|V(G')|=n-1$. By Proposition \\ref{degree n-2}, $\\gamma(G')=2$ and\n$DV_{G'}(v)=|N[\\alpha]| \\le n-3$. Observation \\ref{observation2},\ntogether with $\\gamma(\\langle \\{\\beta\\} \\rangle)=1$ and\n$\\tau(\\langle \\{\\beta\\}\\rangle)=1$, yields $\\gamma(G)=3$ and\n$DV_G(v)=|N[\\alpha]|\\le n-3$. Next suppose that $G$ is connected.\nThen $\\alpha \\beta \\in E(G)$ and $\\alpha$ is a support vertex of\n$G$. (See (B) of Figure \\ref{max1}.) Since $\\Delta(G)<n-1$,\n$\\gamma(G)>1$. Since $\\{v,\\alpha\\}$ is a $\\gamma(G)$-set,\n$\\gamma(G)=2$. Noting that $v$ dominates $V(G)-\\{\\alpha, \\beta\\}$,\nthe number of $\\gamma(G)$-sets containing $v$ equals the number of\nvertices in $G$ that dominates both $\\alpha$ and $\\beta$. Thus $DV_G(v)=2$.\\\\\n\n\\emph{Case 3. There exists a vertex in $N(v)$, say $x$, that is\nadjacent to both $\\alpha$ and $\\beta$:} Notice that $n \\ge 6$ in\nthis case, since $vx, \\alpha x, \\beta x \\in E(G)$ and\n$\\deg_G(v)=\\Delta (G)$ (see (C) of Figure \\ref{max1}). Since $\\{v,\nx\\}$ is a $\\gamma(G)$-set, $\\gamma(G)=2$. If $\\alpha\\beta \\not\\in\nE(G)$, then $DV(v)=|N[\\alpha] \\cap N[\\beta]| \\le n-3$. If\n$\\alpha\\beta \\in E(G)$, then $|N[\\alpha] \\cap N[\\beta]| \\le n-4$\nsince $\\Delta (G)=n-3$. Noting both $\\{v, \\alpha\\}$ and $\\{v,\n\\beta\\}$ are $\\gamma(G)$-sets, we have $DV(v)=2+|N[\\alpha] \\cap N[\\beta]| \\le n-2$.\\\\\n\n\\begin{figure}[htbp]\n\\begin{center}\n\\scalebox{0.5}{\\input{max_n_34.pstex_t}} \\caption{Subcases 4.1\nand 4.2 when $\\Delta(G)=n-3$}\\label{max2}\n\\end{center}\n\\end{figure}\n\n\\emph{Case 4. There exist vertices in $N(v)$ that are adjacent to\n$\\alpha$ and $\\beta$, but no vertex in $N(v)$ is adjacent to both\n$\\alpha$ and $\\beta$:} Let $x_0 \\in N(v) \\cap N(\\alpha)$ and $y_0\n\\in N(v) \\cap N(\\beta)$. We consider two\nsubcases.\\\\\n\n\\emph{Subcase 4.1. $\\alpha \\beta \\not\\in E(G)$ (see (A) of Figure\n\\ref{max2}):} First, assume $\\gamma(G)=2$. This is possible when\n$\\{x_0, y_0\\}$ is a $\\gamma(G)$-set satisfying $N[x_0] \\cup\nN[y_0]=V(G)$. Notice that there's no $\\gamma(G)$-set containing\n$v$ when $\\gamma(G)=2$ since there's no vertex in $G$ that is\nadjacent to both $\\alpha$ and $\\beta$. Thus $DV(v)=0$. Second,\nassume $\\gamma(G)>2$. Since $\\{v, \\alpha, \\beta\\}$ is a\n$\\gamma(G)$-set, $\\gamma(G) =3$. Noticing that every\n$\\gamma(G)$-set contains a vertex in $N[\\alpha]$ and a vertex in\n$N[\\beta]$ and that $N[\\alpha] \\cap N[\\beta] =\\emptyset$, we see\n$$DV(v) = |N[\\alpha]| \\cdot |N[\\beta]| \\le\n\\left(\\frac{|N[\\alpha]|+|N[\\beta]|}{2}\\right)^2 \\le\n\\left(\\frac{n-1}{2}\\right)^2,$$ where the first inequality is the\narithmetic-geometric mean inequality (i.e., $\\frac{a+b}{2} \\geq\n\\sqrt{ab}$ for $a, b \\geq 0$).\\\\\n\n\\emph{Subcase 4.2. $\\alpha \\beta \\in E(G)$ (see (B) of Figure\n\\ref{max2}):} Since $\\{v, \\alpha\\}$ is a $\\gamma(G)$-set,\n$\\gamma(G)=2$. Since there's no vertex in $N(v)$ that is adjacent\nto both $\\alpha$ and $\\beta$, there are only two $\\gamma(G)$-sets\ncontaining $v$, i.e., $\\{v, \\alpha\\}$ and $\\{v, \\beta\\}$. Thus\n$DV(v)=2$. \\hfill\n\\end{proof}\n\n\\noindent\\textbf{Remark.} In the proof of Theorem~\\ref{theorem on\nDelta(G)=n-3}, we observe that one may have $DV(v)=0$ even though\n$\\deg_G(v)=\\Delta(G) \\le n-3$. See Figure \\ref{zeroTDV-max-deg}\nfor a graph of order $n$, $\\deg_G(v)= \\Delta(G)=n-3$,\n$\\gamma(G)=2$, and $DV(v)=0$.\n\n\\begin{figure}[htbp]\n\\begin{center}\n\\scalebox{0.5}{\\input{zeroDV_max_deg.pstex_t}} \\caption{A graph of\norder $9$, $\\deg_G(v)=\\Delta(G)=6$, $DV(v)=0$ with a unique\n$\\gamma$-set $\\{x_0, y_0\\}$}\\label{zeroTDV-max-deg}\n\\end{center}\n\\end{figure}\n\n\n\\section{Domination value on complete $n$-partite graphs}\n\nFor a \\emph{complete} $n$-partite graph $G$, let $V(G)$ be\npartitioned into $n$-partite sets $V_1$, $V_2$, $\\ldots$, $V_n$,\nand let $a_i=|V_i|\\geq 1$ for each $1\\leq i \\leq n$, where $n\\geq\n2$.\n\n\\begin{proposition}\nLet $G=K_{a_1,a_2, \\ldots, a_n}$ be a complete $n$-partite graph\nwith $a_i \\ge 2$ for each $i$ ($1 \\le i \\le n$). Then\n$$\\tau(G)=\\frac{1}{2}\\left[\\left(\\sum_{i=1}^{n} a_i\\right)^2 - \\sum_{i=1}^{n} a_i^2\\right] \\hskip .2in\n\\mbox{ and } \\hskip .2in DV(v)=\\left(\\sum_{i=1}^{n} a_i\\right)-a_j\n\\mbox{ if } v \\in V_j.$$\n\\end{proposition}\n\n\\begin{proof}\nSince $\\Delta(G) < |V(G)|-1$, $\\gamma(G)>1$. Any two vertices from\ndifferent partite sets form a $\\gamma(G)$-set, so $\\gamma(G)=2$.\nIf $v \\in V_j$, then\n\\begin{equation}\\label{DV for n-partite}\nDV(v)=\\deg_G(v)=\\left(\\sum_{i=1}^{n} a_i\\right)-a_j.\n\\end{equation}\nFrom Observation \\ref{observation} and (\\ref{DV for n-partite}),\nwe have\n\\begin{eqnarray*}\n\\sum_{j=1}^{n}\\sum_{v\\in V_j} DV(v) =2 \\tau(G) \\\n\\Longleftrightarrow \\ \\sum_{j=1}^{n}\\left(a_j\\sum_{i=1}^{n} a_i\n-a_{j}^{2}\\right) =2\\tau(G) \\\\ \n \\Longleftrightarrow \\ \\\n\\left(\\sum_{i=1}^{n} a_i\\right)\\left(\\sum_{j=1}^{n}\na_j\\right)-\\sum_{j=1}^{n} a_j^2 =2\\tau(G),\n\\end{eqnarray*}\nand thus the formula for $\\tau(G)$ follows.\\hfill\n\\end{proof}\n\n\\begin{proposition}\\label{n-partite2}\nLet $G=K_{a_1,a_2, \\ldots, a_n}$ be a complete $n$-partite graph\nsuch that $a_i=1$ for some $i$, say $a_j=1$ for $j=1,2,\\ldots, k$,\nwhere $1 \\le k \\le n$. Then $\\tau(G)=k$ and\n\\begin{equation*} DV(v)= \\left\\{\n\\begin{array}{ll}\n1 & \\mbox{ if } v \\in V_j \\  (1 \\le j \\le \\ k)\\\\\n0 & \\mbox{ if } v \\in V_j \\ (k+1 \\le j \\le n) .\n\\end{array} \\right.\n\\end{equation*}\n\\end{proposition}\n\n\\begin{proof}\nSince $\\Delta(G)=|V(G)|-1$, by Observation \\ref{degree n-1},\n$\\gamma(G)=1$ and $DV(v)$ follows. By Observation\n\\ref{observation}, together with $\\gamma(G)=1$, we have $\\tau(G)=\\sum_{v\n\\in V(G)}DV_G(v)=k$.\\hfill\n\\end{proof}\n\n\nIf $a_i=1$ for each $i$ ($1\\leq i\\leq n$), then $G=K_n$. As an\nimmediate consequence of Proposition \\ref{n-partite2}, we have the\nfollowing.\n\n\\begin{corollary}\nIf $G=K_n$ ($n \\ge 1$), then $\\tau(G)= n$ and $DV(v)=1$ for each\n$v \\in V(K_n)$.\n\\end{corollary}\n\n\n\nIf $n=2$, then $G=K_{a_1, a_2}$ is a complete bi-partite graph.\n\\begin{corollary}\nIf $G=K_{a_1, a_2}$, then\n\\begin{equation*}\n\\tau(G)= \\left\\{\n\\begin{array}{ll}\na_1 \\cdot a_2 & \\mbox{ if } a_1, a_2 \\ge 2 \\\\\n2 & \\mbox{ if } a_1=a_2=1\\\\\n1 & \\mbox{ if } \\{a_1, a_2\\} = \\{1,x\\}, \\mbox{ where }x>1 .\n\\end{array} \\right.\n\\end{equation*}\nIf $a_1, a_2 \\ge 2$, then\n\\begin{equation*}\nDV(v)= \\left\\{\n\\begin{array}{ll}\na_2 & \\mbox{ if } v \\in V_1 \\\\\na_1 & \\mbox{ if } v \\in V_2 .\n\\end{array} \\right.\n\\end{equation*}\nIf $a_1=a_2=1$, $DV(v)=1$ for any $v$ in $K_{1,1}$. \nIf $\\{a_1, a_2\\} = \\{1,x\\}$ with $x>1$, say $a_1=1$ and $a_2=x$,\nthen\n\\begin{equation*} DV(v)= \\left\\{\n\\begin{array}{ll}\n1 & \\mbox{ if } v \\in V_1 \\\\\n0 & \\mbox{ if } v \\in V_2 .\n\\end{array} \\right.\n\\end{equation*}\n\\end{corollary}\n\n\n\\section{Domination value on cycles}\n\nLet the vertices of the cycle $C_n$ be labeled $1$ through $n$\nconsecutively in counter-clockwise order, where $n \\ge 3$. Observe\nthat the domination value is constant on the vertices of $C_n$,\nfor each $n$, by vertex-transitivity. Recall that\n$\\gamma(C_n)=\\lceil \\frac{n}{3} \\rceil$ for $n \\ge 3$ (see\np.364, \\cite{CZ}).\\\\\n\n\\noindent {\\bf Examples.} (a) $DM(C_4)=\\{\\{1,\n2\\}, \\{1,3\\}, \\{1,4\\}, \\{2, 3\\}, \\{2, 4\\}, \\{3, 4\\}\\}$ since $\\gamma(C_4)=2$;\nso $\\tau(C_4)=6$ and $DV(i)=3$ for each $i \\in V(C_4)$.\\\\\n\n(b) $\\gamma(C_6)=2$, $DM(C_6)=\\{\\{1, 4\\}, \\{2, 5\\}, \\{3, 6\\} \\}$;\nso $\\tau(C_6)=3$ and $DV(i)=1$ for each $i \\in V(C_6)$.\\\\\n\n\\begin{theorem}\\label{theorem on cycles}\nFor $n \\ge 3$,\n\\begin{equation*}\n\\tau(C_n) = \\left\\{\n\\begin{array}{lr}\n\\ 3 & \\mbox{ if } n \\equiv 0  \\mbox{ (mod 3)} \\ \\\\\n\\ n(1+\\frac{1}{2} \\lfloor \\frac{n}{3} \\rfloor) & \\mbox{ if } n \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\n\\ n & \\mbox{ if } n \\equiv 2  \\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation*}\n\\end{theorem}\n\n\\begin{proof}\nFirst, let $n=3k$, where $k \\ge 1$. Here $\\gamma(C_{n})=k$; a\n$\\gamma(C_{n})$-set $\\Gamma$ comprises $k$ $K_1$'s and $\\Gamma$ is\nfixed by the choice of the first $K_1$. There exists exactly one\n$\\gamma(C_{n})$-set containing the vertex $1$, and there are two\n$\\gamma(C_{n})$-sets omitting the vertex 1 such as $\\Gamma$\ncontaining\nthe vertex 2 and $\\Gamma$ containing the vertex $n$. Thus $\\tau(C_{n})=3$.\\\\\n\nSecond, let $n=3k+1$, where $k \\ge 1$. Here $\\gamma(C_{n})=k+1$; a\n$\\gamma(C_{n})$-set $\\Gamma$ is constituted in exactly one of the\nfollowing two ways: 1) $\\Gamma$ comprises $(k-1)$ $K_1$'s and one\n$K_2$; 2) $\\Gamma$ comprises $(k+1)$ $K_1$'s. \\\\\n\n\\emph{Case 1) $\\langle \\Gamma \\rangle \\cong (k-1)K_1 \\cup K_2$:} Note that\n$\\Gamma$ is fixed by the choice of the single $K_2$. Choosing a\n$K_2$ is the same as choosing its initial vertex\nin the counter-clockwise order. Thus $\\tau=3k+1$.\\\\\n\n\\emph{Case 2) $\\langle \\Gamma \\rangle \\cong (k+1)K_1$:} Note that, since each\n$K_1$ dominates three vertices, there are exactly two vertices,\nsay $x$ and $y$, each of whom is adjacent to two distinct $K_1$'s\nin $\\Gamma$. And $\\Gamma$ is fixed by the placements of $x$ and\n$y$. There are $n=3k+1$ ways of choosing $x$. Consider the\n$P_{3k-2}$ (a sequence of $3k-2$ slots) obtained as a result of\ncutting from $C_{n}$ the $P_3$ centered about $x$. Vertex $y$ may\nbe placed in the first slot of any of the\n$\\lceil\\frac{3k-2}{3}\\rceil=k$ subintervals of the $P_{3k-2}$. As\nthe order of selecting the two vertices\n$x$ and $y$ is immaterial, $\\tau=\\frac{(3k+1)k}{2}$. \\\\\n\nSumming over the two disjoint cases, we get\n$$\\tau(C_{n})=(3k+1)+\\frac{(3k+1)k}{2}=(3k+1)\\left(1+\\frac{k}{2}\\right)= n \\left(1+\\frac{1}{2} \\left\\lfloor\\frac{n}{3} \\right\\rfloor \\right).$$\n\nFinally, let $n=3k+2$, where $k \\ge 1$. Here $\\gamma(C_{n})=k+1$;\na $\\gamma(C_{n})$-set $\\Gamma$ comprises of only $K_1$'s and is\nfixed by the placement of the only vertex which is adjacent to two\ndistinct $K_1$'s in $\\Gamma$. Thus $\\tau(C_{n})=n$. \\hfill\n\\end{proof}\n\n\\begin{corollary}\nLet $v\\in V(C_n)$, where $n \\ge 3$. Then\n\\begin{equation*}\nDV(v) = \\left\\{\n\\begin{array}{lr}\n\\ 1 & \\mbox{ if } n \\equiv 0  \\mbox{ (mod 3)} \\ \\\\\n\\ \\frac{1}{2} \\lceil\\frac{n}{3}\\rceil (1+ \\lceil\\frac{n}{3}\\rceil) & \\mbox{ if } n \\equiv 1  \\mbox{ (mod 3)} \\ \\\\\n\\ \\lceil\\frac{n}{3}\\rceil & \\mbox{ if } n \\equiv 2 \\mbox{ (mod 3) }.\n\\end{array} \\right.\n\\end{equation*}\n\\end{corollary}\n\n\\begin{proof}\nIt follows by Observation~\\ref{observation}, Observation~\\ref{observation1}, and Theorem~\\ref{theorem on cycles}.~\\hfill\n\\end{proof}\n\n\n\n\\section{Domination value on paths}\n\nLet the vertices of the path $P_n$ be labeled $1$ through $n$\nconsecutively. Recall that $\\gamma(P_n)=\\lceil \\frac{n}{3} \\rceil$ for $n \\ge 2$.\\\\\n\n\n\\noindent {\\bf Examples.} (a) $\\gamma(P_4)=2$, $DM(P_4)=\\{\n\\{1,3\\}, \\{1,4\\}, \\{2, 3\\}, \\{2,4\\}\\}$; so $\\tau(P_4)=4$ and\n$DV(i)=2$ for each $i \\in V(P_4)$.\\\\\n\n(b) $\\gamma_t(P_5)=2$, $DM(P_5)=\\{\\{1, 4\\}, \\{2,4\\}, \\{2,5\\}\\}$;\nso $\\tau(P_5)=3$, and\n\\begin{equation*}\nDV(i)= \\left\\{\n\\begin{array}{ll}\n1 & \\mbox{ if } i =1,5\\\\\n2 & \\mbox{ if } i=2,4\\\\\n0 & \\mbox{ if } i =3 .\n\\end{array} \\right.\n\\end{equation*}\n\n\\noindent\\textbf{Remark.} Since $P_n \\subset C_n$ with the same vertex set,\nby Proposition \\ref{subgraph-tau}, we have $\\tau(P_n) \\le\n\\tau(C_n)$ for $n \\ge 3$, as one can verify from the theorem\nbelow.\n\n\\begin{theorem}\\label{theorem on paths}\nFor $n \\ge 2$,\n\\begin{equation*} \\tau(P_n) = \\left\\{\n\\begin{array}{lr}\n\\ 1 & \\mbox{ if } n \\equiv 0  \\mbox{ (mod 3)} \\ \\\\\n\\ n+ \\frac{1}{2}\\lfloor \\frac{n}{3} \\rfloor (\\lfloor \\frac{n}{3}\\rfloor-1) & \\mbox{ if } n \\equiv 1  \\mbox{ (mod 3)} \\ \\\\\n\\ 2+ \\lfloor \\frac{n}{3} \\rfloor & \\mbox{ if } n \\equiv 2 \\mbox{\n(mod 3)} .\n\\end{array} \\right.\n\\end{equation*}\n\\end{theorem}\n\n\\begin{proof}\nFirst, let $n=3k$, where $k \\ge 1$. Then $\\gamma(P_n)=k$ and a\n$\\gamma(P_n)$-set $\\Gamma$ comprises $k$ $K_1$'s. In this case,\neach vertex in $\\Gamma$ dominates three vertices, and no vertex of\n$P_{n}$ is dominated by more than one vertex. Thus none of the\nend-vertices of $P_n$ belongs to any $\\Gamma$, which contains and\nis fixed by the vertex $2$; hence $\\tau(P_n)=1$.\\\\\n\nSecond, let $n=3k+1$, where $k \\ge 1$. Here $\\gamma(P_n)=k+1$; a\n$\\gamma(P_n)$-set $\\Gamma$ is constituted in exactly one of the\nfollowing two ways: 1) $\\Gamma$ comprises $(k-1)$ $K_1$'s and one\n$K_2$; 2) $\\Gamma$ comprises $(k+1)$ $K_1$'s.\\\\\n\n\\emph{Case 1) $\\langle \\Gamma \\rangle \\cong (k-1)K_1 \\cup K_2$, where $k \\ge\n1$:} Note that $\\Gamma$ is fixed by the placement of the single\n$K_2$, and none of the end-vertices belong to any $\\Gamma$, as\neach component with cardinality $c$ in $\\langle \\Gamma \\rangle$ dominates $c+2$\nvertices. Initial vertex of $K_2$ may be placed in one of the $n\n\\equiv 2$ (mod $3$) slots. Thus $\\tau=k$.\\\\\n\n\\emph{Case 2) $\\langle \\Gamma \\rangle \\cong (k+1)K_1$, where $k \\ge 1$:} A\n$\\Gamma$ containing both end-vertices of the path is unique (no\nvertex is doubly dominated). The number of $\\Gamma$ containing\nexactly one of the end-vertices (one doubly dominated vertex) is\n$2{k \\choose 1}=2k$. The number of $\\Gamma$ containing none of the\nend-vertices (two doubly dominated vertices) is ${k \\choose 2}=\n\\frac{k(k-1)}{2}$. Thus\n$\\tau=1+2k+\\frac{k(k-1)}{2}$.\\\\\n\nSumming over the two disjoint cases, we get\n$$\\tau(P_n)=k+\\left(1+2k+\\frac{k(k-1)}{2} \\right)=3k+1+ \\frac{k(k-1)}{2}=\nn+\\frac{1}{2} \\left\\lfloor\\frac{n}{3}\\right\\rfloor\n\\left(\\left\\lfloor \\frac{n}{3}\\right\\rfloor-1 \\right).$$\n\nFinally, let $n=3k+2$, where $k \\ge 0$. Here $\\gamma(P_n)=k+1$,\nand $\\gamma(P_n)$-set $\\Gamma$ comprises of ($k+1$) $K_1$'s. Note\nthat there's no $\\Gamma$ containing both end-vertices of $P_n$.\nThe number of $\\Gamma$ containing exactly one of the end-vertices\n(no doubly dominated vertex) of the path is two.  The number of\n$\\Gamma$ containing neither of the end-vertices (one doubly\ndominated vertex) is $k$. Summing the two disjoint cases, we have\n$\\tau(P_n)=2+k=2+ \\lfloor \\frac{n}{3}\\rfloor$. \\hfill\n\\end{proof}\n\nFor the domination value of a vertex on $P_n$, note that\n$DV(v)=DV(n+1-v)$ for $1 \\le v \\le n$ as $P_n$ admits the obvious\nautomorphism carrying $v$ to $n+1-v$. More precisely, we have the\nclassification result which follows. First, as an immediate\nconsequence of Theorem \\ref{theorem on paths}, we have the\nfollowing result.\n\n\\begin{corollary} \\label{path on 3k}\nLet $v \\in V(P_{3k})$, where $k \\ge 1$. Then\n\\begin{equation*}\nDV(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0,1  \\mbox{ (mod 3)} \\\\\n1 & \\mbox{ if } v \\equiv 2  \\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation*}\n\\end{corollary}\n\n\\begin{proposition}\nLet $v \\in V(P_{3k+1})$, where $k \\ge 1$. Write $v=3q+r$, where\n$q$ and $r$ are non-negative integers such that $0 \\le r < 3$.\nThen, noting $\\tau(P_{3k+1})=\\frac{1}{2}(k^2+5k+2)$, we have\n\\begin{equation*} DV(v)= \\left\\{\n\\begin{array}{ll}\n\\frac{1}{2}q(q+3) & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3) } \\ \\\\\n(q+1)(k-q+1) & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3) } \\ \\\\\n\\frac{1}{2}(k-q)(k-q+3) & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3) }.\n\\end{array} \\right.\n\\end{equation*}\n\\end{proposition}\n\n\\begin{proof}\nLet $\\Gamma$ be a $\\gamma(P_{3k+1})$-set for $k \\ge 1$. We\nconsider two cases.\\\\\n\n\\textit{Case 1) $\\langle \\Gamma \\rangle \\cong (k-1)K_1 \\cup K_2$, where $k \\ge\n1$:} Denote by $DV^1(v)$ the number of such $\\Gamma$'s containing\n$v$. Noting $\\tau=k$ in this case, we have\n\\begin{equation}\\label{path 3k+1 case1}\nDV^1(v)= \\left\\{\n\\begin{array}{ll}\nq & \\mbox{ if } v \\equiv 0  \\mbox{ (mod 3)} \\ \\\\\n0 & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\nk-q & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation}\nWe prove by induction on $k$. One can easily check (\\ref{path 3k+1\ncase1}) for $k=1$. Assume that (\\ref{path 3k+1 case1}) holds for\n$G=P_{3k+1}$ and consider $G'=P_{3k+4}$. First, notice that each\n$\\Gamma$ of the $k$ $\\gamma(P_{3k+1})$-sets of $G$ induces a\n$\\gamma(P_{3k+4})$-set $\\Gamma'=\\Gamma \\cup \\{3k+3\\}$ of $G'$.\nAdditionally, $G'$ has the $\\gamma(P_{3k+4})$-set $\\Gamma^*$ that\ncontains and is determined by $\\{3k+2, 3k+3\\}$, which does not\ncome from any $\\gamma(P_{3k+1})$-set of $G$. The presence of\n$\\Gamma^*$ implies that $DV^1_{G'}(v)=DV^1_G(v)+1$ for $v \\equiv\n2$ (mod 3), where $v \\le 3k+1$. Clearly, $DV^1_{G'}(3k+2)=1$,\n$DV^1_{G'}(3k+3)=k+1$, and $DV^1_{G'}(3k+4)=0$.\\\\\n\n\n\\textit{Case 2) $\\langle \\Gamma \\rangle \\cong (k+1)K_1$, where $k \\ge 1$:}\nDenote by $DV^2(v)$ the number of such $\\Gamma$'s containing $v$.\nFirst, suppose both end-vertices belong to the unique $\\Gamma$ and\ndenote by $DV^{2,1}(v)$ the number of such $\\Gamma$'s containing\n$v$. Then we have\n\\begin{equation}\\label{Case 3-1}\nDV^{2,1}(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0, 2  \\mbox{ (mod 3)} \\ \\\\\n1 & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation}\n\nSecond, suppose exactly one end-vertex belongs to each $\\Gamma$;\ndenote by $DV^{2,2}(v)$ the number of such $\\Gamma$'s containing\n$v$. Then, noting $\\tau=2k$ in this case, we have\n\\begin{equation}\\label{Case 3-2}\nDV^{2,2}(v)= \\left\\{\n\\begin{array}{ll}\nq & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3)} \\ \\\\\nk & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\nk-q & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation}\nWe prove by induction on $k$. One can easily check (\\ref{Case\n3-2}) for $k=1$. Assume that (\\ref{Case 3-2}) holds for\n$G=P_{3k+1}$ and consider $G'=P_{3k+4}$. First, notice that each\n$\\Gamma$ of the $k$ $\\gamma(P_{3k+1})$-sets of $G$ containing the\nleft end-vertex $1$ induces a $\\gamma(P_{3k+4})$-set\n$\\Gamma'=\\Gamma \\cup \\{3k+3\\}$ of $G'$. Second, each $\\Gamma$ of\n$k$ $\\gamma(P_{3k+1})$-sets of $G$ containing the right end-vertex\n$3k+1$ induces a $\\gamma(P_{3k+4})$-set $\\Gamma'=\\Gamma \\cup\n\\{3k+4\\}$ of $G'$. Third, a $\\gamma(P_{3k+1})$-set $\\Gamma$ of $G$\ncontaining $1$ and $3k+1$ (both left and right end-vertices of\n$G$) induces a $\\gamma(P_{3k+4})$-set $\\Gamma^{*1}=\\Gamma \\cup\n\\{3k+3\\}$ of $G'$ (making $3k+2$ the only doubly dominated vertex\nin $G'$). Additionally, $\\Gamma^{*2}=\\{v \\in V(P_{3k+1}) \\mid v \\equiv 2 \\mbox{ (mod 3)}\\} \\cup \\{3k+2, 3k+4\\}$ is a\n$\\gamma(P_{3k+4})$-set for $G'$, which does not come from any\n$\\gamma(P_{3k+1})$-set of $G$. The presence of $\\Gamma^{*1}$ and\n$\\Gamma^{*2}$ imply that\n\\begin{equation*}\nDV^{2,2}_{G'}(v)= \\left\\{\n\\begin{array}{ll}\nDV^{2,2}_G(v) & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3)} \\ \\\\\nDV^{2,2}_G(v)+1 & \\mbox{ if } v \\equiv 1,2 \\mbox{ (mod 3)}\n\\end{array} \\right.\n\\end{equation*}\nfor $v \\le 3k+1$.\nClearly, $DV^{2,2}_{G'}(3k+2)=1$, $DV^{2,2}_{G'}(3k+3)=k+1$, and $DV^{2,2}_{G'}(3k+4)=k+1$.\\\\\n\n\nThird, suppose no end-vertex belongs to $\\Gamma$; denote by\n$DV^{2,3}(v)$ the number of such $\\Gamma$'s containing $v$. Then,\nnoting $\\tau= {k\\choose 2}$ in this case and setting ${a\\choose\nb}=0$ when $a<b$, we have\n\\begin{equation} \\label{Case 3-3}\nDV^{2,3}(v)= \\left\\{\n\\begin{array}{ll}\n\\frac{1}{2}(q-1)q & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3)} \\ \\\\\nq(k-q) & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\n\\frac{1}{2}(k-q-1)(k-q) & \\mbox{ if } v \\equiv 2\\mbox{ (mod 3)} .\n\\end{array} \\right.\n\\end{equation}\nAgain, we prove by induction on $k$. Since $DV^{2,3}(v)=0$ for\neach $v \\in V(P_{4})$, we consider $k \\ge 2$. One can easily check\n(\\ref{Case 3-3}) for the base, $k=2$. Assume that (\\ref{Case 3-3})\nholds for $G=P_{3k+1}$ and consider $G'=P_{3k+4}$, where $k \\ge\n2$. First, notice that each $\\Gamma$ of the $k\\choose 2$\n$\\gamma(P_{3k+1})$-sets of $G$ containing neither end-vertices of\n$G$ induces a $\\gamma(P_{3k+4})$-set $\\Gamma'=\\Gamma \\cup\n\\{3k+3\\}$ of $G'$. Additionally, each $\\Gamma_r$ of the $k$\n$\\gamma(P_{3k+1})$-sets of $G$ containing the right-end vertex\n$3k+1$ of $G$ induces a $\\gamma(P_{3k+4})$-set $\\Gamma_r'=\\Gamma_r\n\\cup \\{3k+3\\}$ of $G'$ (making $3k+2$ one of the two\ndoubly-dominated vertices in $G'$): If we denote by $DV^r_G(v)$\nthe number of such $\\Gamma_r$'s containing $v$ in $G$, then one\ncan readily check\n\\begin{equation*}\nDV^r_G(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3)} \\ \\\\\nq & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\nk-q & \\mbox{ if } v \\equiv 2\\mbox{ (mod 3)} ,\n\\end{array} \\right.\n\\end{equation*}\nagain by induction on $k$. Thus, the presence of $\\Gamma'_r$\nimplies $DV_{G'}^{2,3}(v)=DV_G^{2,3}(v)+DV^r_G(v)$ for $v \\le\n3k+1$. Clearly, $DV^{2,3}_{G'}(3k+2)=0=DV^{2,3}_{G'}(3k+4)$\nand $DV^{2,3}_{G'}(3k+3)= {k\\choose 2}+k=\\frac{1}{2}k(k+1)$.\\\\\n\nSumming over the three disjoint cases (\\ref{Case 3-1}), (\\ref{Case\n3-2}), and (\\ref{Case 3-3}) for $\\langle \\Gamma \\rangle \\cong (k+1)K_1$, we have\n\\begin{equation}\\label{path 3k+1 case3}\nDV^2(v)= \\left\\{\n\\begin{array}{ll}\nq + \\frac{1}{2}(q-1)q& \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3)} \\ \\\\\n1+k+q(k-q) & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3)} \\ \\\\\nk-q+ \\frac{1}{2}(k-q-1)(k-q)& \\mbox{ if } v \\equiv 2 \\mbox{ (mod\n3)} .\n\\end{array} \\right.\n\\end{equation}\nNow, by summing over (\\ref{path 3k+1 case1}) and (\\ref{path 3k+1\ncase3}), i.e., $DV(v)=DV^1(v)+DV^2(v)$, we obtain the formula\nclaimed in this proposition.~\\hfill\n\\end{proof}\n\n\n\\begin{proposition}\nLet $v \\in V(P_{3k+2})$, where $k \\ge 0$. Write $v=3q+r$, where\n$q$ and $r$ are non-negative integers such that $0 \\le r < 3$.\nThen, noting $\\tau(P_{3k+2})=k+2$, we have\n\\begin{equation*}\nDV(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3) } \\ \\\\\n1+q & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3) } \\ \\\\\nk+1-q & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3) } .\n\\end{array} \\right.\n\\end{equation*}\n\\end{proposition}\n\n\\begin{proof}\nLet $\\Gamma$ be a $\\gamma(P_{3k+2})$-set for $k \\ge 0$. Then $\\langle\n\\Gamma \\rangle \\cong (k+1) K_1$. Note that no $\\Gamma$ contains both\nend-vertices of $P_{3k+2}$.\\\\\n\nFirst, suppose $\\Gamma$ contains\nexactly one end-vertex, and denote by $DV'(v)$ the number of such\n$\\Gamma$'s containing $v$. Noting $\\tau=2$ in this case, for $v\n\\in V(P_{3k+2})$, we have\n\\begin{equation}\\label{path on 3k+2 case1}\nDV'(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3) } \\ \\\\\n1 & \\mbox{ if } v \\equiv 1,2 \\mbox{ (mod 3) }.\n\\end{array} \\right.\n\\end{equation}\n\nNext, suppose $\\Gamma$ contains no end-vertices (thus $k \\ge 1$),\nand denote by $DV''(v)$ the number of such $\\Gamma$'s containing\n$v$. Noting $\\tau=k$ in this case, we have\n\\begin{equation}\\label{path on 3k+2 case2}\nDV''(v)= \\left\\{\n\\begin{array}{ll}\n0 & \\mbox{ if } v \\equiv 0 \\mbox{ (mod 3) } \\ \\\\\nq & \\mbox{ if } v \\equiv 1 \\mbox{ (mod 3) } \\ \\\\\nk-q & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3) } .\n\\end{array} \\right.\n\\end{equation}\nWe prove by induction on $k$. One can easily check (\\ref{path on\n3k+2 case2}) for the base, $k=1$. Assume that (\\ref{path on 3k+2\ncase2}) holds for $G=P_{3k+2}$ and consider $G'=P_{3k+5}$. First,\nnotice that each $\\Gamma$ of the $k$ $\\gamma(P_{3k+2})$-sets\ncontaining neither end-vertex of $G$ induces a\n$\\gamma(P_{3k+5})$-set $\\Gamma'=\\Gamma \\cup \\{3k+4\\}$.\nAdditionally, the only $\\gamma(P_{3k+2})$-set $\\Gamma$ of $G$\ncontaining the right-end vertex $3k+2$ of $G$ induces a\n$\\gamma(P_{3k+5})$-set $\\Gamma^\\star=\\Gamma \\cup \\{3k+4\\}$ of $G'$\n(making $3k+3$ the only doubly-dominated vertex). The presence of\n$\\Gamma^\\star$ implies that\n\\begin{equation*}\nDV_{G'}''(v)= \\left\\{\n\\begin{array}{ll}\nDV_G''(v) & \\mbox{ if } v \\equiv 0, 1 \\mbox{ (mod 3) } \\\\\nDV_G''(v)+1 & \\mbox{ if } v \\equiv 2 \\mbox{ (mod 3) }\n\\end{array} \\right.\n\\end{equation*}\nfor $v \\le 3k+2$. Clearly, $DV_{G'}''(3k+3)=0=DV_{G'}''(3k+5)$ and\n$DV_{G'}''(3k+4)=k+1$.\\\\\n\nNow, by summing over the two disjoint cases (\\ref{path on 3k+2\ncase1}) and (\\ref{path on 3k+2 case2}), i.e.,\n$DV(v)=DV'(v)+DV''(v)$, we obtain the formula claimed in this\nproposition. \\hfill\n\\end{proof}\n\n\\textit{Acknowledgement.} The author thanks Cong X. Kang for\nsuggesting the concept of \\emph{domination value} and his valuable\ncomments and suggestions. The author also thanks the referee for a couple of helpful comments.\n\n\n\\nocite{*}\n\\bibliographystyle{amsplain}\n", "meta": {"timestamp": "2012-03-02T02:03:38", "yymm": "1109", "arxiv_id": "1109.6277", "language": "en", "url": "https://arxiv.org/abs/1109.6277", "abstract": "A set $D \\subseteq V(G)$ is a \\emph{dominating set} of $G$ if every vertex not in $D$ is adjacent to at least one vertex in $D$. A dominating set of $G$ of minimum cardinality is called a $\\gamma(G)$-set. For each vertex $v \\in V(G)$, we define the \\emph{domination value} of $v$ to be the number of $\\gamma(G)$-sets to which $v$ belongs. In this paper, we study some basic properties of the domination value function, thus initiating \\emph{a local study of domination} in graphs. Further, we characterize domination value for the Petersen graph, complete $n$-partite graphs, cycles, and paths.", "subjects": "Combinatorics (math.CO)", "title": "Domination Value in Graphs", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9814534316905262, "lm_q2_score": 0.8633916205190225, "lm_q1q2_score": 0.8473786688512392}}
{"url": "https://arxiv.org/abs/1510.04471", "title": "Nearest points and delta convex functions in Banach spaces", "abstract": "Given a closed set $C$ in a Banach space $(X, \\|\\cdot\\|)$, a point $x\\in X$ is said to have a nearest point in $C$ if there exists $z\\in C$ such that $d_C(x) =\\|x-z\\|$, where $d_C$ is the distance of $x$ from $C$. We shortly survey the problem of studying how large is the set of points in $X$ which have nearest points in $C$. We then discuss the topic of delta-convex functions and how it is related to finding nearest points.", "text": "\\section{Nearest points in Banach spaces}\n \n\\subsection{Background}\\label{sec intro}\n \nLet $(X, \\|\\cdot\\|)$ be a real Banach space, and let $C\\subseteq X$ be a non-empty closed set. Given $x\\in X$, its distance from $C$ is given by\n\\[d_C(x) = \\inf_{y\\in C} \\|x-y\\|.\\]\nIf there exists $z\\in C$ with $d_C(x) = \\|x-z\\|$, we say that $x$ has a \\emph{nearest point} in $C$. Let also\n\\[N(C) = \\big\\{x\\in X: x \\text{ has a nearest point in $C$ }\\big\\}.\\]\nOne can then ask questions about the structure of the set $N(C)$. This question has been studied in \\cite{Ste63, Lau78, Kon80, Zaj83, BF89, DMP91, Dud04, RZ11, RZ12} to name just a few. More specifically, the following questions are at the heart of this note:\n\n\\medskip\n\\begin{center}\n\\emph{Given a nonempty closed set $C\\subseteq X$, how large is the set $N(C)$? When is it non-empty?}\n\\end{center}\n\\medskip\n\nOne way to do so is to consider sets which are large in the set theoretic sense, such as dense $G_{\\delta}$ sets. We begin with a few definitions.\n\n\\begin{defin}\nIf $N(C) =  X$, i.e., every point in $X$ has a nearest point in $C$, then $C$ is said to be proximinal. If $N(C)$ contains a dense $G_{\\delta}$ set, then $C$ is said to be almost proximinal.\n\\end{defin}\n\nIn passing we recall that If every point in $X$ is uniquely proximinal then $C$ is said to be a \\emph{Chebyshev set}  It has been conjectured for over half a century,  that in Hilbert space \nChebyshev sets are necessarily convex, but this is only proven for weakly closed sets \\cite{BV10}. See also \\cite{FM15} for a recent survey on the topic.\n\nFor example, closed convex sets in reflexive spaces are proximinal, as well as closed sets in finite dimensional spaces. See \\cite{BF89}. One can also consider stronger notions of ``large\" sets. See Section \\ref{sec porous}. First, we also need the following definition.\n\n\\begin{defin}\nA Banach space is said to be a (sequentially) Kadec space if for each sequence $\\{x_n\\}$ that converges weakly to $x$ with $\\lim \\|x_n\\| = \\|x\\|$, $\\{x_n\\}$ converges to $x$ in norm, i.e.,\n\\[\\lim_{n\\to \\infty} \\|x-x_n\\| =0.\\]\n\\end{defin}\n\nWith the above definitions in hand, the following result holds.\n\n\\begin{thm}[Lau \\cite{Lau78}, Borwein-Fitzpatrick \\cite{BF89}]\\label{thm Lau BF}\nIf $X$ is a reflexive Kadec space and $C\\subseteq X$ is closed, then $C$ is almost proximinal.\n\\end{thm}\n\nThe assumptions on $X$ are in fact necessary.\n\n\\begin{thm}[Konjagin \\cite{Kon80}]\\label{thm Kon}\nIf $X$ is not both Kadec and reflexive, then there exist $C\\subseteq X$ closed and $U\\subseteq X\\setminus C$ open such that no $x\\in U$ has a nearest point in $C$.\n\\end{thm}\n\nIt is known that under stronger assumption on $X$ one can obtain stronger results on the set $N(C)$. See Section \\ref{sec porous}.\n \n\n\\medskip\n  \n\n\\subsection{Fr\\'echet sub-differentiability and nearest points}\\label{sec diff}\n \nWe begin with a definition.\n\n\\begin{defin}\nAssume that $f: X\\to \\mathbb R$ is a real valued function with $f(x)$ finite. Then $f$ is said to be Fr\\'echet sub-differentiable at $x\\in X$ if there exists $x^*\\in X^*$ such that\n\\begin{align}\\label{cond subdiff}\n\\liminf_{y\\to 0}\\frac{f(x+y)-f(x)-x^*(y)}{\\|y\\|} \\ge 0.\n\\end{align}\nThe set of points in $X^*$ that satisfy \\eqref{cond subdiff} is denoted by $\\partial f(x)$.\n\\end{defin} \n\nSub-derivatives have been found to have many applications in approximation theory. See for example \\cite{BF89, BZ05, BL06, BV10, Pen13}. \n \nOne of the connections between sub-differentiability and the nearest point problem was studied in \\cite{BF89}. Given $C\\subseteq X$ closed, the following modification of a construction of \\cite{Lau78} was introduced.\n\\begin{align*}\nL_n(C) = \\Big\\{x\\in X\\setminus C : \\exists x^*\\in \\mathbb S_{X^*} \\text{ s.t. }\\sup_{\\delta>0}~\\inf_{z\\in C\\cap B(x,d_C(x)+\\delta)}~x^*(x-z) > \\big(1-2^{-n}\\big)d_C(x)\\Big\\},\n\\end{align*}\nwhere $\\mathbb S_{X^*}$ denotes the unit sphere of $X^*$. Also, let\n\\begin{align*}\nL(C) = \\bigcap_{n=1}^{\\infty}L_n(C).\n\\end{align*}\nThe following is known.\n\\begin{prop}[Borwein-Fitzpatrick \\cite{BF89}]\\label{prop open}\nFor every $n\\in \\mathbb N$, $L_n(C)$ is open. In particular, $L(C)$ is $G_{\\delta}$.\n\\end{prop}\nFinally, let\n\\begin{align*}\n\\Omega(C)  = & \\Big\\{ x\\in X\\setminus C : \\exists x^*\\in \\mathbb S_{X^*}, \\text{ s.t. } \\forall \\epsilon >0, \\exists \\delta>0,\n\\\\ & ~~\\quad\\inf_{z\\in C\\cap B(x,d_C(x)+\\delta)}x^*(x-z) > \\big(1-\\epsilon\\big)d_C(x)\\Big\\}.\n\\end{align*}\n\nWhile $L(C)$ is $G_{\\delta}$ by Proposition \\ref{prop open}, under the assumption that $X$ is reflexive, the following is known. \n\n\n\\begin{prop}[Borwein-Fitzpatrick \\cite{BF89}]\nIf $X$ is reflexive then $\\Omega(C) = L(C)$. In particular, $\\Omega(C)$ is $G_{\\delta}$.\n\\end{prop}\n\nThe connection to sub-differentiability is given in the following proposition.\n\n\\begin{prop}[Borwein-Fitzpatrick \\cite{BF89}]\nIf $x\\in X\\setminus C$ and $\\partial d_C(x) \\neq \\emptyset$, then $x\\in \\Omega(C)$.\n\\end{prop}\n\nAlso, the following result is known.\n\n\\begin{thm}[Borwein-Preiss \\cite{BP87}]\\label{thm var}\nIf $f$ is lower semicontiuous on a reflexive Banach space, then $f$ is Fr\\'echet sub-differentiable on a dense set.\n\\end{thm}\n\nIn fact, Theorem \\ref{thm var} holds under a weaker assumption. See \\cite{BP87, BF89}.\nSince the distance function is lower semicontinuous, it follows that it is sub-differentiable on a dense subset, and therefore, by the above propositions, $\\Omega(C)$ is a dense $G_{\\delta}$ set. Thus, in order to prove Theorem \\ref{thm Lau BF}, it is only left to show that every $x\\in \\Omega(C)$ has a nearest point in $C$. Indeed, if $\\{z_n\\}\\subseteq C$ is a minimizing sequence, then by extracting a subsequence, assume that $\\{z_n\\}$ has a weak limit $z\\in C$. By the definition of $\\Omega(C)$, there exists $x^* \\in \\mathbb S_{X^*}$ such that\n\\[\\|x-z\\| \\ge x^*(x-z) = \\lim_{n\\to \\infty} x^*(x-z_n) \\ge d_C(x) = \\lim_{n\\to \\infty}\\|x-z_n\\|.\\]\nOn the other hand, by weak lower semicontinuity of the norm,\n\\[\\lim_{n\\to \\infty} \\|x-z_n\\| \\ge \\|x-z\\|,\\]\nand so $\\|x-z\\| = \\lim\\|x-z_n\\|$. Since it is known that $\\{z_n\\}$ converges weakly to $z$, the Kadec property implies that in fact $\\{z_n\\}$ converges in norm to $z$. Thus $z$ is a nearest point. This completes the proof of Theorem \\ref{thm Lau BF}.\n\nThis scheme of proof from \\cite{BF89} shows that differentiation arguments can be used to prove that $N(C)$ is large.\n\n\\medskip\n\n\\subsection{Nearest points in non-Kadec spaces}\\label{sec non-Kadec}\nIt was previously mentioned that closed convex sets in reflexive spaces are proximinal. It also known that non-empty ``Swiss cheese\" sets (sets whose complement is a mutually disjoint union of open convex sets) in reflexive spaces are almost proximinal \\cite{BF89}. These two examples show that for some classes of closed sets, the Kadec property can be removed. Moreover, one can consider another, weaker, way to ``measure\" whether a set $C\\subseteq X$ has ``many\" nearest points: ask whether the set of nearest points in $C$ to points in $X\\setminus C$ is dense in the boundary of $C$. Note that if $C$ is almost proximinal, then nearest points are dense in the boundary. The converse, however, is not true. In \\cite{BF89} an example of a non-Kadec reflexive space was constructed where for every closed set, the set of nearest points is dense in its boundary. The following general question is still open.\n\n\\begin{qst}\nLet $(X, \\|\\cdot\\|)$ be a reflexive Banach space and $C\\subseteq X$ closed. Is the set of nearest points in $C$ to points in $X\\setminus C$ dense in its boundary?\n\\end{qst}\n\n Relatedly,  if the set $C$ is norm closed and bounded in a space with the \\emph{Radon-Nikodym property} as is the caae of reflexive space, then $N(C)$ is nonempty and is large enough so that $\\overline{\\rm conv} C = \\overline{\\rm conv} N(C)$ \\cite{BF89}.\n\n\n\\medskip\n \n\\subsection{Porosity and nearest points}\\label{sec porous}\n \nAs was mentioned in subsection \\ref{sec diff}, one can consider stronger notions of ``large\" sets. One is the following notion.\n \n\\begin{defin}\nA set $S\\subseteq X$ is said to be porous if there exists $c\\in (0,1)$ such that for every $x\\in X$ and every $\\epsilon>0$, there is a $y \\in B(0,\\epsilon)\\setminus \\{0\\}$ such that\n\\[ B(x+y, c\\|y\\|) \\cap S = \\emptyset.\\]\nA set is said to be $\\sigma$-porous if it a countable union of porous sets. Here and in what follows, $B(x,r)$ denotes the closed ball around $x$ with radius $r$.\n\\end{defin}\n\nSee \\cite{Zaj05, LPT12} for a more detailed discussion on porous sets. It is known that every $\\sigma$-porous set is of the first category, i.e., union of nowhere dense set. Moreover, it is known that the class of $\\sigma$-porous sets is a proper sub-class of the class of first category sets. When $X=\\mathbb R^n$, one can show that every $\\sigma$-porous set has Lebesgue measure zero. This is not the case for every first category set: $\\mathbb R$ can be written as a disjoint union of a set of the first category and a set of Lebesgue measure zero. Hence, the notion of porosity automatically gives a stronger notion of large sets: every set whose complement is $\\sigma$-porous is also a dense $G_{\\delta}$ set.\n\n\\medskip\n\nA Banach space $(X, \\|\\cdot\\|)$ is said to be \\emph{uniformly convex} if the function\n\\begin{align}\\label{def uni conv}\n\\delta(\\epsilon) = \\inf\\left\\{ 1-\\left\\|\\frac{x+y}2\\right\\| ~ : ~ x, y\\in \\mathbb S_X, \\|x-y\\| \\ge \\epsilon \\right\\},\n\\end{align}\nis strictly positive whenever $\\epsilon>0$. Here $\\mathbb S_X$ denotes the unit sphere of $X$. In \\cite{DMP91} the following was shown.\n \n\\begin{thm}[De Blasi-Myjak-Papini \\cite{DMP91}]\\label{thm DMP}\nIf $X$ is uniformly convex, then $N(C)$ has a $\\sigma$-porous compliment.\n\\end{thm}\n\nIn fact, \\cite{DMP91} proved a stronger result, namely that for every $x$ outside a $\\sigma$-porous set, the minimization problem is \\emph{well posed}, i.e., there is unique minimizer to which every minimizing sequence converges. See also \\cite{FP91, RZ11, RZ12} for closely related results in this direction. \n\nThe proof of Theorem \\ref{thm DMP} builds on ideas developed in \\cite{Ste63}. However, it would be interesting to know whether one could use differentiation arguments as in Section \\ref{sec diff}. This raises the following question:\n \n \n\\begin{qst}\nCan differentiation arguments be used to give an alternative proof of Theorem \\ref{thm DMP}?\n\\end{qst}\n\nMore specifically, if one can show that $\\partial d_C \\neq \\emptyset$ outside a $\\sigma$-porous set, then by the arguments presented in Section \\ref{sec diff}, it would follow that $N(C)$ has a $\\sigma$-porous complement. Next, we mention two important results regarding differentiation in Banach spaces. \n \n\\begin{thm}[Preiss-Zaj\\'i\\v{c}ek \\cite{PZ84}]\\label{thm PZ}\nIf $X$ has a separable dual and $f:X\\to \\mathbb R$ is continuous and convex, then $X$ is Fr\\'echet differentiable outside a $\\sigma$-porous set.\n\\end{thm}\n \nSee also \\cite[Sec. 3.3]{LPT12}. Theorem \\ref{thm PZ} implies that if, for example, $d_C$ is a linear combination of convex functions (see more on this in Section \\ref{sec DC}), then $N(C)$ has a $\\sigma$-porous complement. Also, we have the following.\n\n\\begin{thm}[C\\'uth-Rmoutil \\cite{CR13}]\\label{thm CR}\nIf $X$ has a separable dual and $f:X\\to \\mathbb R$ is Lipschitz, then the set of points where $f$ is Fr\\'echet sub-differentiable but not differentiable is $\\sigma$-porous.\n\\end{thm}\n\nSince $d_C$ is 1-Lipschitz, the questions of seeking points of sub-differentiability or points of differentiability are similar. Theorem \\ref{thm PZ} and Theorem \\ref{thm CR} remain true if we consider $f:A\\to \\mathbb R$ where $A\\subseteq X$ is open and convex.\n \n\n\\bigskip\n \n \n\\section{DC functions and DC sets}\\label{sec DC}\n \n\\subsection{Background}\n\n\\begin{defin}\nA function $f:X\\to \\mathbb R$ is said to be delta-convex, or DC, if it can be written as a difference of two convex functions on $X$. \n\\end{defin}\n\nThis notion was introduced in \\cite{Har59} and was later studied by many authors. See for example \\cite{KM90, Cep98, Dud01, VZ01, DVZ03, BZ05, Pav05, BB11}. In particular, \\cite{BB11} gives a good introduction to this topic. We will discuss here only the parts that are closely related to the nearest point problem.\n \nThe following is an important proposition. See for example \\cite{VZ89, HPT00} for a proof.\n \n\\begin{prop}\\label{prop select}\nIf $f_1,\\dots,f_k$ are DC functions and $f:X\\to \\mathbb R$ is continuous and $f(x)\\in \\big\\{f_1(x),\\dots,f_n(x)\\big\\}$. Then $f$ is also DC.\n\\end{prop}\n\nThe result is true if we replace the domain $X$ by any convex subset.\n\n\\medskip\n\n\\subsection{DC functions and nearest points}\n\nShowing that a given function is in fact DC is a powerful tool, as it allows us to use many known results about convex and DC functions. For example, if a function is DC on a Banach space with a separable dual, then by Theorem \\ref{thm PZ}, it is differentiable outside a $\\sigma$-porous set. In the context of the nearest point problem, if we know that the distance function is DC, then using the scheme presented in Section \\ref{sec diff}, it would follow that $N(C)$ has a $\\sigma$-porous complement. The same holds if we have a difference of a convex function and, say, a smooth function.\n\n\\medskip\n \nThe simplest and best known example is when $(X,\\|\\cdot\\|)$ is a Hilbert space, where we have the following.\n \n\\begin{align*}\nd_C^2(x) & = \\inf_{y\\in C}\\|x-y\\|^2\n\\\\ & = \\inf_{y\\in C}\\Big[\\|x\\|^2-2\\langle x,y\\rangle +\\|y\\|^2\\Big]\n\\\\ & = \\|x\\|^2 - 2\\sup_{y\\in C}\\Big[\\langle x,y\\rangle - \\|y\\|^2/2\\Big],\n\\end{align*}\nand the function $x\\mapsto  \\sup_{y\\in C}\\Big[\\langle x,y\\rangle - \\|y\\|^2/2\\Big]$ is convex as a supremum of affine functions. Hence $d_C^2$ is DC on $X$.\nMoreover, in a Hilbert space we have the following result (see \\cite[Sec. 5.3]{BZ05}).\n\n\\begin{thm}\\label{thm local DC}\nIf $(X, \\|\\cdot\\|)$ is a Hilbert space, $d_C$ is locally DC on $X\\setminus C$.\n\\end{thm}\n\n\\begin{proof}\nFix $y\\in C$ and $x_0\\in X\\setminus C$. It can be shown that if we let $f_y(x) = \\|x-y\\|$, then $f_y$ satisfies\n\\begin{align*}\n\\big\\|f_y'(x_1)-f_y'(x_2)\\big\\|_{X^*} \\le L_{x_0}\\|x_1-x_2\\|, ~~ x_1,x_2 \\in B_{x_0},\n\\end{align*}\nwhere $L_{x_0} = \\frac 4 {d_S(x_0)}$ and $B_{x_0} = B\\Big(x_0, \\frac 1 2 d_C(x_0)\\Big)$. In particular,\n\\begin{align}\\label{lip prop}\n\\big(f_y'(x+tv_1)-f_y'(x+t_2v)\\big)(v) \\le L_{x_0}(t_2-t_1), ~~ v\\in \\mathbb S_X, t_2> t_1 \\ge 0,\n\\end{align}\nwhenever $x+t_1v, x+t_2v \\in B_{x_0}$. Next, the convex function $F(x) = \\frac {L_{x_0}} 2 \\|x\\|^2$ satisfies\n\\begin{align}\\label{anti lip hilbert}\n\\big(F'(x_1)-F'(x_2)\\big)(x_1-x_2) \\ge L_{x_0}\\|x_1-x_2\\|^2, ~~ \\forall x_1,x_2\\in X.\n\\end{align}\nIn particular\n\\begin{align}\\label{anti lip}\n\\big(F'(x+t_2v)-F'(x+t_1v)\\big)(v) \\ge L_{x_0}(t_2-t_1),  ~~ v\\in \\mathbb S_X, ~t_2>t_1\\ge 0.\n\\end{align}\nAltogether, if $g_y(x) = F(x)-f_y(x)$, then\n\\begin{align*}\n\\big(g_y'(x+t_2v)-g_y'(x+t_1v)\\big)(v) \\stackrel{\\eqref{lip prop}\\wedge \\eqref{anti lip}}{\\ge} 0, ~~ v\\in \\mathbb S_X, ~t_2>t_1\\ge 0,\n\\end{align*}\nwhenever $x+t_1v, x+t_2v \\in B_{x_0}$. This implies that $g_y$ is convex on $B_{x_0}$. It then follows that\n\\begin{align*}\nd_C(x) & = \\frac {L_{x_0}} 2 \\|x\\|^2 -\\sup_{y\\in C}\\Bigg[~\\frac {L_{x_0}} 2 \\|x\\|^2-\\|x-y\\|\\Bigg] = h(x) - \\sup_{y\\in C}g_y(x)\n\\end{align*}\nis DC on $B_{x_0}$. \n\\end{proof}\n\n\\begin{remark}\nEven in $\\mathbb R^2$ there are sets for which $d_C$ is not DC everywhere (not even locally DC), as was shown in \\cite{BB11}. Thus, the most one could hope for is a locally DC function on $X\\setminus C$.\n\\end{remark}\n\n\\medskip\n\nGiven $q\\in (0,1]$, a norm $\\|\\cdot\\|$ is said to be $q$\\emph{-H\\\"older smooth} at a point $x\\in X$ if there exists a constant $K_x\\in (0,\\infty)$ such that for every $y\\in \\mathbb S_X$ and every $\\tau>0$,\n\\begin{align*}\n\\frac{\\|x+\\tau y\\|}{2} + \\frac{\\|x-\\tau y\\|}{2} \\le 1+K_x\\tau^{1+q}.\n\\end{align*}\nIf $q=1$ then $(X,\\|\\cdot\\|)$ is said to be \\emph{Lipschitz smooth} at $x$.\nThe spaces $L_p$, $p \\ge 2$ are known to be Lipschitz smooth, and in general $L_p$, $p>1$, is $s$-H\\\"older smooth with $s = \\min\\{1,p-1\\}$. \n\n\\medskip\n\nA Banach space is said to be $p$\\emph{-uniformly convex} if for every $x,y\\in \\mathbb S_X$,\n\\begin{align*}\n1-\\left\\|\\frac{x+y}{2}\\right\\| \\ge L\\|x-y\\|^p.\n\\end{align*}\nNote that this is similar to assuming that $\\delta(\\epsilon) = L\\epsilon^p$ in \\eqref{def uni conv}. The spaces $L_p$, $p>1$, are $r$-uniformly convex with $r = \\max\\{2,p\\}$. \n \nOne could ask whether the scheme of proof of Theorem \\ref{thm local DC} can be used in a more general setting.\n\n\\begin{prop}\\label{prop Hilbert}\nLet $(X, \\|\\cdot\\|)$ be a Banach space, $C\\subseteq X$ a closed set, and fix $x_0\\in X\\setminus C$ and $y\\in C$. Assume that there exists $r_0$ such that $f_y(x) = \\|x-y\\|$ has a Lipschitz derivative on $B(x_0,r_0)$:\n\\begin{align}\\label{lip of der}\n\\big\\|f_y'(x_1)-f_y'(x_2)\\| \\le L_{x_0}\\|x_1-x_2\\|.\n\\end{align}\nThen the norm is Lipschitz smooth on $-y + B_{x_0} = B(x_0-y, r_0)$. If in addition there exists a function $F: X \\to \\mathbb R$ satisfying \n\\begin{align}\\label{strong conv}\n\\big(F'(x_1)-F'(x_2)\\big)(x_1-x_2) \\ge L_{x_0}\\|x_1-x_2\\|^2, ~~ \\forall x_1,x_2\\in B(x_0,r_0),\n\\end{align}\nthen $(X,\\|\\cdot\\|)$ admits an equivalent norm which is 2-uniformly convex. In particular, if $X=L_p$ then $p=2$.\n\\end{prop}\n\n\\begin{proof}\nTo prove the first assertion note that \\eqref{lip of der} is equivalent to \n\\begin{align*}\n\\|x-y+h\\|+\\|x-y-h\\|-2\\|x-y\\| \\le L_{x_0}\\|h\\|^2, ~~ x\\in B_{x_0}.\n\\end{align*}\nSee for example \\cite[Prop. 2.1]{Fab85}.\n\nTo prove the second assertion, note that a function that satisfies \\eqref{strong conv} is also known as \\emph{strongly convex}: one can show that \\eqref{strong conv} is in fact equivalent to the condition\n\\begin{align*}\nf\\left(\\frac{x_1+x_2}{2}\\right) \\le \\frac 1 2f(x_1)+\\frac 1 2 f(x_2) - C\\|x_1-x_2\\|^2,\n\\end{align*}\nfor some constant $C$. See for example \\cite[App. A]{SS07}. This implies that there exists an equivalent norm which is 2-uniformly convex (\\cite[Thm 5.4.3]{BV10}).\n\n\\end{proof}\n\n\\begin{remark}\nFrom \\cite{Ara88} it is know that if $F:X\\to \\mathbb R$ satisfies \n\\begin{align*}\n\\big(F'(x_1)-F'(x_2)\\big)(v) \\ge L\\|x_1-x_2\\|^2,\n\\end{align*}\nfor \\emph{all} $x_1,x_2\\in X$, and also that $F$ is twice (Fr\\'echet) differentiable at one point, then $(X,\\|\\cdot\\|)$ is isomorphic to a Hilbert space.\n\\end{remark}\n\n\\begin{remark}\nIf we replace the Lipschitz condition by a H\\\"older condition\n\\begin{align*}\n\\big\\|f_y'(x_1)-f_y'(x_2)\\big\\| \\le \\|x_1-x_2\\|^{\\beta}, ~~ \\beta <1, \n\\end{align*}\nthen in order to follow the same scheme of proof of Theorem \\ref{thm local DC}, instead of \\eqref{anti lip hilbert}, we would need a function $F$ satisfying\n\\begin{align*}\n\\big(F'(x_1)-F'(x_2)\\big)(x_1-x_2) \\ge \\|x_1-x_2\\|^{1+\\beta}, ~~ x_1,x_2\\in B_{x_0}.\n\\end{align*}\nwhich implies\n\\begin{align}\\label{anti holder}\n\\big\\|F'(x_1)-F'(x_2)\\big\\| \\ge \\|x_1-x_2\\|^{\\beta}, ~~ x_1,x_2\\in B_{x_0}.\n\\end{align}\nIf $G= (F')^{-1}$, then we get\n\\begin{align*}\n\\|Gx_1-Gx_2\\| \\le \\|x_1-x_2\\|^{1/\\beta}, ~~ x_1,x_2 \\in F'(B_{x_0}),\n\\end{align*}\nwhich can occur only if $G$ is a constant. Hence \\eqref{anti holder} cannot hold and the scheme of proof cannot be used if we replace the Lipschitz condition by a H\\\"older condition.\n\\end{remark}\n\n\n\\medskip\n \n\\subsection{DC sets, DC representable sets}\n \n\\begin{defin}\nA set $C$ is is said to be a DC set if $C=A\\setminus B$ where $A,B$ are convex.\n\\end{defin} \nWe can also consider the following class of sets.\n\\begin{defin}\nA set $C\\subseteq X$ is said to be DC representable if there exists a DC function $f:X\\to R$ such that $C = \\big\\{x\\in X: f(x) \\le 0\\big\\}$.\n\\end{defin}\nNote that if $C = A\\setminus B$ is a DC set, then we can write $C = \\Big\\{\\mathbbm 1_{B}-\\mathbbm 1_{A}+1/2 \\le 0\\Big\\}$, where $\\mathbbm 1_A$, $\\mathbbm 1_B$ are the indicator functions of $A,B,$ respectively.  Therefore, $C$ is DC representable. Moreover, we have the following.\n\n\\begin{thm}[Thach \\cite{Th93}]\nAssume that $X$ and $Y$ are two Banach space, and $T:Y\\to X$ is surjective map with $\\mathrm{ker}(T) \\neq \\emptyset$. Then for any set $M\\subseteq X$ there exists a DC representable set $D\\subseteq Y$, such that $M= T(D)$.\n\\end{thm}\n\nAlso, the following is known. See \\cite{HPT00}.\n \n\\begin{prop}\nIf $C$ is a DC representable set, then there exist $A,B \\subseteq X\\oplus \\mathbb R$ convex, such that $x\\in C \\iff (x,x') \\in A\\setminus B$.\n\\end{prop}\n \n\\begin{proof}\nDefine $g_1(x,x') = f_1(x)-x'$, $g_2(x,x') = f_2(x)-x'$. Let $A = \\big\\{(x,x') : g_1(x,x')\\le 0\\big\\}$, $B = \\big\\{(x,x'): g_2(x,x') \\le 0\\big\\}$. Then $x\\in C \\iff (x,x') \\in A\\setminus B$.\n\\end{proof}\n\nIn particular, every DC representable set in $X$ is a projection of a DC set in $X\\oplus \\mathbb R$. The following theorem was proved in \\cite{TK96}\n \n\\begin{thm}[Thach-Konno \\cite{TK96}]\nIf $X$ is a reflexive Banach space and $C\\subseteq X$ is closed, then $C$ is DC representable.\n\\end{thm}\n\nThis raises the following question.\n \n\\begin{qst}\\label{qst DC}\nIs it true that for some classes of spaces, e.g. uniformly convex spaces, there exists $\\alpha>0$ such that $d_C^{\\alpha}$ is locally DC on $X\\setminus C$ whenever $C$ is a DC representable set?\n\\end{qst}\nIf the answer to Question \\ref{qst DC} is positive, then by the discussion in subsection \\ref{sec diff} we could conclude that $N(C)$ has a $\\sigma$-porous complement, thus giving an alternative proof of Theorem \\ref{thm DMP}. One could also ask Question \\ref{qst DC} for DC sets instead of DC representable sets.\n\n\\medskip\n\nTo end this note, we discuss some simple cases where DC and DC representable sets can be used to study the nearest point problem. \n\n\n\\begin{prop}\nAssume that $C = X \\setminus \\bigcup_{a\\in \\Lambda}U_a$, where each $U_a$ is an open convex set. Then $d_C$ is locally DC (in fact, locally concave) on $X\\setminus C$.\n\\end{prop}\n\n\\begin{proof}\nFirst, it is shown in \\cite[Sec. 3]{BF89} that if $a\\in \\Lambda$, then $d_{X\\setminus U_a}$ is concave on $U_a$. Next, it also shown in \\cite{BF89} that if $x\\in U_a$ then $d_{X\\setminus U_a}(x) = d_C(x)$. In particular, $d_C$ is concave on $U_a$. \n\\end{proof}\n\n\\begin{prop}\nAssume that  $C = A\\setminus B$ is a closed DC set, and assume $A$ is closed and $B$ is open, then $d_C$ is convex whenever $d_{C}(x) \\le d_{A\\cap B}$.\n\\end{prop}\n \n\\begin{proof}\nSince $A = \\big(A\\setminus B\\big) \\bigcup B$, we have\n$$d_A(x) = \\min\\big\\{d_{A\\setminus B}(x), d_{A\\cap B}(x)\\big\\} = \\min\\big\\{d_{C}(x), d_{A\\cap B}(x)\\big\\}.$$\nHence, if $d_C(x) \\le d_{A\\cap B}(x)$ then $d_C(x) = d_A(x)$ is convex.\n\\end{proof}\n \n\\begin{prop}\nAssume that $C$ is a DC representable set, i.e., $C = \\big\\{x\\in X: f_1(x)-f_2(x)\\le 0\\big\\}$, and that $f_2(x) =\\max_{1\\le i \\le m}\\varphi_i(x)$, where $\\varphi_i$ is affine. Then $d_C$ is DC on $X$.\n\\end{prop}\n \n\\begin{proof}\nWrite\n\\begin{align*}\nC & = \\Big\\{x : f_1(x)-f_2(x)\\le 0\\Big\\}\n\\\\ &= \\Big\\{x  :  f_1(x)-\\max_{1\\le i \\le m}\\varphi_i(x) \\le 0\\Big\\}\n\\\\ & = \\Big\\{x  : \\min_{1\\le i \\le m}\\big(f_1(x)-\\varphi_i(x)\\big) \\le 0\\Big\\}\n\\\\ & = \\bigcup_{i=1}^n \\Big\\{x : f_1(x)-\\varphi_i(x) \\le 0\\Big\\}.\n\\end{align*}\nwhere the sets $\\Big\\{x~ : ~ f_1(x)-\\varphi_i(x) \\le 0\\Big\\}$ are convex sets. Hence, we have that\n\\[d_C(x) = \\min_{1\\le i \\le m}d_{C_i}(x),\\]\nis a minimum of convex sets and therefore by Proposition \\ref{prop select} is a DC function.\n\\end{proof}\n\nIn \\cite{Cep98} it was shown that if $X$ is superreflexive, then any Lipschitz map is a uniform limit of DC functions. See also \\cite[Sec. 5.1]{BV10}. We have the following simple result.\n\n\\begin{prop}\nIf $X$ is separable, then $d_C$ is a limit (not necessarily uniform) of DC functions.\n\\end{prop} \n\n\\begin{proof}\nIf $X$ is separable, i.e., there exists a countable $Q = \\{q_1,q_2,\\dots\\}\\subseteq X$ with $\\bar Q = X$. We have\n\\begin{align*}\nd_C(x) = \\inf_{z\\in C}\\|x-z\\| = \\inf_{z\\in C\\cap Q}\\|x-z\\| = \\lim_{n\\to \\infty} \\Big[\\min_{z\\in C\\cap Q_n}\\|x-z\\|\\Big],\n\\end{align*}\nwhere $Q_n = \\{q_1,q_2,\\dots,q_n\\}$. Again by Proposition \\ref{prop select} we have that $\\min_{z\\in C\\cap Q_n}\\|x-z\\|$ is a DC function as a minimum of convex functions. \n\\end{proof}\n\n\\section{Conclusion}\nDespite many decades of study, the core questions addressed in this note are still far from settled. We hope that our analysis will encourage others to take up the quest, and also to reconsider the related \\emph{Chebshev problem} \\cite{B07,BV10}.\n\n \n", "meta": {"timestamp": "2015-10-16T02:08:58", "yymm": "1510", "arxiv_id": "1510.04471", "language": "en", "url": "https://arxiv.org/abs/1510.04471", "abstract": "Given a closed set $C$ in a Banach space $(X, \\|\\cdot\\|)$, a point $x\\in X$ is said to have a nearest point in $C$ if there exists $z\\in C$ such that $d_C(x) =\\|x-z\\|$, where $d_C$ is the distance of $x$ from $C$. We shortly survey the problem of studying how large is the set of points in $X$ which have nearest points in $C$. We then discuss the topic of delta-convex functions and how it is related to finding nearest points.", "subjects": "Functional Analysis (math.FA)", "title": "Nearest points and delta convex functions in Banach spaces", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9899864281950695, "lm_q2_score": 0.8558511414521923, "lm_q1q2_score": 0.8472810145929291}}
{"url": "https://arxiv.org/abs/2004.08316", "title": "On Zeckendorf Related Partitions Using the Lucas Sequence", "abstract": "Zeckendorf proved that every positive integer has a unique partition as a sum of non-consecutive Fibonacci numbers. Similarly, every natural number can be partitioned into a sum of non-consecutive terms of the Lucas sequence, although such partitions need not be unique. In this paper, we prove that a natural number can have at most two distinct non-consecutive partitions in the Lucas sequence, find all positive integers with a fixed term in their partition, and calculate the limiting value of the proportion of natural numbers that are not uniquely partitioned into the sum of non-consecutive terms in the Lucas sequence.", "text": "\\section{Introduction} \nThe Fibonacci numbers have fascinated mathematicians for centuries with many interesting properties. By convention, the Fibonacci sequence $\\left\\{F_n\\right\\}_{n=0}^{\\infty}$ is defined as follows: let $F_0 = 0$, $F_1 = 1$, and $F_n = F_{n-1} + F_{n-2}$, for $n\\ge 2$.\nA beautiful theorem of Zeckendorf \\cite{Z} states that every natural number $n$ can be uniquely written as a sum of non-consecutive Fibonacci numbers. This gives the so-called Zeckendorf partition of $n$. A formal statement of Zeckendorf's theorem is as follows:\n\\begin{thm}[Zeckendorf]\\label{p1}\nFor any $n\\in\\mathbb{N}$, there exists a unique increasing sequence of positive integers $\\{c_1, c_2, \\ldots, c_k\\}$ such that $c_1\\ge 2$, $c_i\\ge c_{i-1}+2$ for $i = 2, 3, \\ldots, k$, and $n = \\sum_{i=1}^kF_{c_i}$. \n\\end{thm}\nMuch work has been done to understand the structure of Zeckendorf partitions and their applications (see \\cite{BDEMMTW1, BDEMMTW2, B, CSH, Fr, MG1, MG2, HS, L, MW1, MW2}) and to generalize them (see \\cite{Chu, Day, DDKMMV, DFFHMPP, FGNPT, GTNP, Ho, K, ML, MMMS, MMMMS}). In this paper, we study the partition of natural numbers into Lucas numbers. The Lucas sequence $\\left\\{L_n\\right\\}_{n=0}^{\\infty}$ is defined as follows: let $L_0 = 2$, $L_1 = 1$, and $L_n = L_{n-1} + L_{n-2}$, for $n\\ge 2$. As the Lucas sequence is closely related to the Fibonacci sequence, it is not surprising that we can also partition natural numbers using Lucas numbers. \n\n\\begin{thm}[Zeckendorf]\\label{p2}\n    Every natural number can be partitioned into the sum of non-consecutive terms of the Lucas sequence.\n\\end{thm}\n\nNote that the distinction between Theorems \\ref{p1} and \\ref{p2} lies in the \\textit{uniqueness} property of such partitions of natural numbers in the Fibonacci and Lucas sequences.~Although $5$ is uniquely partitioned into $F_5 = 5$ in $\\{F_2, F_3, \\ldots \\}$, its partition is not unique in the Lucas sequence as $5 = L_0 + L_2 = 2 + 3$ and $5 = L_1 + L_3 = 1 + 4$. In \\cite{B2}, Brown shows various ways to have a unique partition using Lucas sequence. \\footnote{For more on Brown's criteria, see \\cite{BHLMT1, BHLMT2}.} In this paper, we prove the following results.\n\n\\begin{thm}\\label{m1}\nIf we allow $L_0$ and $L_2$ to appear simultaneously in a partition, each natural number can have at most two distinct non-consecutive partitions in the Lucas sequence.\n\\end{thm}\n\n\n\n\\begin{thm}\\label{m2} Suppose that we do not allow $L_0$ and $L_2$ to appear simultaneously in a partition. \nThe set of all positive integers having the summand $L_k$ in their partition is given by \n$$Z(k) \\ =\\ \\begin{cases}\\left\\{2+3n+\\left\\lfloor\\frac{n+1}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}, & {\\rm if}~ k = 0,\\\\\n\\left\\{3n+\\left\\lfloor\\frac{n+\\Phi^2}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}, &{\\rm if}~ k = 1,\\\\\n\\left\\{L_k\\left\\lfloor\\frac{n+\\Phi^2}{\\Phi}\\right\\rfloor+ nL_{k+1}+j\\, :\\, n\\ge 0 ~{\\rm and}~  0\\le j\\le L_{k-1}-1\\right\\}, &{\\rm if}~ k \\ge 2.\n\\end{cases}$$\n\\end{thm}\nTheorem \\ref{m2} is an analogue of \\cite[Theorem 3.4]{MG1}. For $k\\ge 0$, we find all positive integers having the summand $L_k$ in their partition. We have a different formula when $k = 0$ instead of one formula for all values of $k$ as in \\cite[Theorem 3.4]{MG1}. \n\nOur next result is predicted by \\cite[Theorem 1]{Cha}, which deals with general recurrence relations; however, in the case of Lucas numbers, we can relate Lucas partitions to the golden string. \n\n\n\\begin{thm}\\label{m3}\nIf we allow $L_0$ and $L_2$ to appear simultaneously in a partition, the proportion of natural numbers that are not uniquely partitioned into the sum of non-consecutive terms of the Lucas sequence converges to $\\frac{1}{3\\Phi+1}$, where $\\Phi$ is the golden ratio. \n\\end{thm}\n\n\n\\section{Preliminaries}\n\\subsection{Definitions}\n\\begin{defi}\nLet $A = \\{a_0, a_1, \\ldots, a_m \\}$ be the set consisting of the first $m+1$ terms of the sequence $\\big\\{a_k\\big\\}_{k=0}^{\\infty}$.~We say a proper subset $B$ of $A$ is a \\textit{non-consecutive subset} of $A$ if the elements of $B$ are pairwise non-consecutive in $\\big\\{a_k\\big\\}_{k=0}^{\\infty}$.~Furthermore, we say a sum $S$ is a \\textit{non-consecutive sum} of $A$ if $S$ is the sum of distinct elements of $A$ that are pairwise non-consecutive in $\\big\\{a_k\\big\\}_{k=0}^{\\infty}$. \n\\end{defi}\n\\begin{defi}\nLet $A_m = \\{L_0, L_1, \\ldots, L_m \\}$ denote the set consisting of the first $m+1$ terms of the Lucas sequence.\n\\end{defi}\n\n\n\n\n\n\\subsection{The golden string}\nThe golden string $S = BABBABABBABBA\\ldots$ is defined to be the infinite string of $A$'s and $B$'s constructed recursively as follows. Let $S_1 = A$ and $S_2 = B$, and then, for $k\\ge 3$, $S_k$ is the concatenation of $S_{k-1}$ and $S_{k-2}$, which we denote by $S_{k-1}\\circ S_{k-2}$. For example, $S_3 = S_2\\circ S_1 = a_2\\circ a_1 = BA$, $S_4 = S_3\\circ S_2 = a_2a_1\\circ a_2 = BAB$, $S_5  = S_4\\circ S_3  = BABBA$, and so on. Interestingly, the golden string is highly connected to the Zeckendorf partition \\cite{MG2}. As we will see later, the string is also closely related to the partitions of natural numbers into Lucas numbers. \n\\begin{rek}\\label{r1}\\normalfont We mention two properties of the golden string that we will use in due course. \n\\begin{itemize}\n\\item[(1)] For $j\\ge 1$, the $(F_{2j})$th character of $S$ is $B$ and the $(F_{2j+1})$th character of $S$ is $A$. This can be easily proved using induction. \n\\item[(2)] The number of $B$'s amongst the first $n$ characters of $S$ is given by \n$\\left\\lfloor\\frac{n+1}{\\Phi}\\right\\rfloor$,\nwhere $\\Phi = \\frac{1+\\sqrt{5}}{2}$ is the golden ratio. For a proof of this result, see \\cite[Lemma 3.3]{MG2}.\n\\end{itemize}\n\\end{rek}\n\n\\section{At Most Two Partitions}\n\nIn this section, we present our results that determine the maximum number of non-consecutive partitions that a natural number can have in the Lucas sequence, the proofs of which are adapted from \\cite{HR}. Before we prove Theorem \\ref{m1}, we introduce the following preliminary lemmas. For the proofs of Lemmas \\ref{Lemma3} and \\ref{Lemma5}, see Appendix B.\n\n\\begin{lem}\\label{Lemma3}\nLet $S$ be any non-consecutive sum of $A_m$.~Then\n    \\begin{enumerate}\n            \\item if $m$ is odd, $S$ assumes all values from 0 to $L_{m+1}-1$ inclusive, and\n            \\item if $m$ is even, then $S$ assumes all values from 0 to $L_{m+1}+1$ inclusive, excluding $L_{m+1}$.\n    \\end{enumerate}\n\\end{lem}\n\n\n\n\n\\begin{lem}\\label{Lemma5}\nIf $m \\geq 0$, then $L_{2m+1}+1$ has exactly two non-consecutive partitions in the Lucas sequence.\n\\end{lem}\n\n\n\n\n\\begin{proof}[Proof of Theorem \\ref{m1}]\nIt suffices to show that for every non-negative integer $m$, there is no natural number that is equal to three or more distinct non-consecutive sums of $A_m$.~We proceed by strong induction.~No natural is equal to three or more distinct non-consecutive sums of $A_0$ and $A_1$.~This shows the base case.~Assume Theorem \\ref{m1} holds for all non-negative integers less than or equal to $m=k$.~In our first case, suppose that $k$ is odd.~From Lemma \\ref{Lemma3}, the non-consecutive sums that we can form from $A_k$ are the values from 0 to $L_{k+1}-1$ inclusive.~Hence, when we add the term $L_{k+1}$ to $A_k$, all new non-consecutive sums that can be formed must be at least $L_{k+1}$.~This implies there is no possible way in which we can form a third distinct non-consecutive sum of $A_{k+1}$ for any natural number because there is no intersection between the non-consecutive sums in which we can form before and after the addition of the term $L_{k+1}$.~When $k \\geq 2$ is even, we have from Lemma \\ref{Lemma3} that all non-consecutive sums we can form from $A_k$ are the values from 0 to $L_{k+1}+1$ inclusive, excluding $L_{k+1}$.~When we add the term $L_{k+1}$ to $A_k$, all new non-consecutive sums that can be formed are at least $L_{k+1}$ with $L_{k+1}+1$ being the only non-consecutive sum formed again, namely $L_{k+1}+L_1$.~By Lemma \\ref{Lemma5}, we know that $L_{k+1}+1$ has exactly two distinct non-consecutive partitions in the Lucas sequence.~Therefore, there is no possible way in which we can form a third distinct non-consecutive sum of $A_{k+1}$ for any natural number.~This completes the inductive step.\n\\end{proof}\n\n\n\n\n\n\n\\section{Partitions with a Fixed Term}\n\n\nLet $\\mathcal{X}_k$ denote the set of all positive integers having $L_k$ as the smallest summand in their partition. Let $\\mathcal{Q}_k = (q_k(j))_{j\\ge 1}$ be the strictly increasing sequence obtained by rearranging the elements of $\\mathcal{X}_k$ into ascending numerical order. We consider the cases $k = 0$ and $k\\ge 1$ separately. \n\n\n\\subsection{When $k = 0$}\n\nTable 1 replaces each term $q_k(j)$ in $\\mathcal{Q}_k$ with an ordered list of the summands in its partition. \n\n\\begin{center}\n\\begin{tabular}{cccccccccccc}\n Row  &               &          &     &        &    &         &     &        &  &          &    \\\\\n \\hline\n1     &               & $L_0$    &     &        &    &         &     &        &  &         &     \\\\\n2     &               & $L_0$    &     & $L_3$  &    &         &     &        &  &         &    \\\\\n3     &               & $L_0$    &     &        &    & $L_4$   &     &        &  &         &     \\\\\n$4$   &               & $L_0$    &     &        &    &         &     & $L_5$  &  &         &     \\\\\n$5$   &               & $L_0$    &     & $L_3$  &    &         &     & $L_5$  &  &         &       \\\\\n$6$   &               & $L_0$    &     &        &    &         &     &        &  &  $L_6$  &   \\\\\n$7$   &               & $L_0$    &     & $L_3$  &    &         &     &        &  &  $L_6$  &     \\\\\n$8$   &               & $L_0$    &     &        &    & $L_4$   &     &        &  &  $L_6$  &    \n\\end{tabular}\n\\end{center}\n\n\\begin{center}\nTable 1. The partitions of the positive integers having $L_0$ as their smallest summand.  \n\\end{center}\n\n\\begin{lem}\\label{l1}\nFor $j\\ge 3$, the rows of Table 1 for which $L_{j}$ is the largest summand are those numbered from $F_{j-1}+1$ to $F_{j}$ inclusive. \n\\end{lem}\n\n\\begin{proof}\nWe prove by induction. \\textit{Base cases:} it is easy to check that the statement of the lemma is true for $j = 3$ and $j = 4$. \\textit{Inductive hypothesis:} assume that it is true for all $j$ such that $3\\le j\\le m$ for some $m\\ge 4$. By the inductive hypothesis, the number of rows such that their largest summands are no greater than $L_{m-1}$ is\n$$1+\\sum_{j=3}^{m-1}(F_{j} - F_{j-1}) \\ =\\ F_{m-1},$$\nwhich is also the number of rows whose largest summand is $L_{m+1}$. Due to the inductive hypothesis, the rows whose largest summand is $L_m$ are numbered from $F_{m-1}+1$ to $F_{m}$ inclusive. Therefore, the rows whose largest summand is $L_{m+1}$ are numbered from $F_{m}+1$ to $F_{m+1}$, as desired. This completes our proof. \n\\end{proof}\n\n\\begin{lem}\\label{k1}\nFor $j\\ge 1$, we have\n$$q_k(j+1)-q_k(j) \\ =\\ \\begin{cases}L_{2}, & \\mbox{ if }A \\mbox{ is the }j\\mbox{th character of }S,\\\\\nL_{3}, &\\mbox{ if }B \\mbox{ is the }j\\mbox{th character of }S.\\end{cases}$$\n\\end{lem}\n\n\\begin{proof}\nWe prove by induction. \\textit{Base cases:} it is easy to check that the statement of the lemma is true for $1\\le j\\le F_4-1$. \\textit{Inductive hypothesis:} suppose that it is true for $1\\le j\\le F_m-1$ for some $m\\ge 4$. By Lemma \\ref{l1}, the number of rows in Table 1 whose largest summand is no greater than $L_{m-1}$ is \n$$1+ \\sum_{j=3}^{m-1}(F_{j} - F_{j-1}) \\ =\\ F_{m-1},$$\nwhich is also the number of rows whose largest summand is $L_{m+1}$. Furthermore, the rows for which $L_{m+1}$ is the largest summand are numbered from $F_{m}+1$ to $F_{m+1}$ inclusive. Therefore, the ordering of the rows in Table 1 implies that \n$q_k(i+F_{m})\\ =\\ q_k(i) + L_{m+1}$,\nfor $1\\le i\\le F_{m-1}$. Hence, for $1\\le i\\le F_{m-1}-1$, we have \n\\begin{align*}\n    q_k(i+1+F_{m}) - q_k(i+F_{m}) &\\ =\\ (q_k(i+1) + L_{m+1})-(q_k(i) + L_{m+1})\\ =\\ q_k(i+1) - q_k(i).\n\\end{align*}\nBy the construction of $S$, the substring comprising of its first $F_{m-1}$ characters is identical to the substring of its characters numbered from $F_{m}+1$ to $F_{m+1}$ inclusive. Thus the lemma is true for $F_{m}+1\\le j\\le F_{m+1}-1$. It remains to show that it is true for $j = F_{m}$. We have\n\\begin{align*}\n    q_k(F_m+1) - q_k(F_m) \\ =\\begin{cases}\\ L_{m+1} - (L_{m}+L_{m-2} + \\cdots + L_4) \\ =\\ L_3,&\\mbox{ if }m\\mbox{ is even,}\\\\\n    L_{m+1} - (L_m + L_{m-2} + \\cdots + L_3)\\ =\\ L_2,&\\mbox{ if }m\\mbox{ is odd.}\\end{cases}\n\\end{align*}\nBy Remark \\ref{r1} item (1), we know that the lemma is true for $j = F_{m}$, completing the proof.  \n\\end{proof}\n\n\\subsection{When $k\\ge 1$}\n\nTable 2 replaces each term $q_k(j)$ in $\\mathcal{Q}_k$ with an ordered list of the summands in its partition. \n\\begin{center}\n\\begin{tabular}{cccccccccccc}\n Row  &               &          &     &        &    &         &     &        &  &          &    \\\\\n \\hline\n1     &               & $L_k$    &     &        &    &         &     &        &  &         &     \\\\\n2     &               & $L_k$    &     & $L_{k+2}$  &    &         &     &        &  &         &    \\\\\n3     &               & $L_k$    &     &        &    & $L_{k+3}$   &     &        &  &         &     \\\\\n$4$   &               & $L_k$    &     &        &    &         &     & $L_{k+4}$  &  &         &     \\\\\n$5$   &               & $L_k$    &     & $L_{k+2}$  &    &         &     & $L_{k+4}$  &  &         &       \\\\\n$6$   &               & $L_k$    &     &        &    &         &     &        &  &  $L_{k+5}$  &   \\\\\n$7$   &               & $L_k$    &     & $L_{k+2}$  &    &         &     &        &  &  $L_{k+5}$  &     \\\\\n$8$   &               & $L_k$    &     &        &    & $L_{k+3}$   &     &        &  &  $L_{k+5}$  &     \n\\end{tabular}\n\\end{center}\n\\begin{center}\nTable 2. The partitions of the positive integers having $L_k$ as their smallest summand.  \n\\end{center}\n\nTable 2 is similar to Table 1 in \\cite{MG1}. The next lemma follows from \\cite[Lemma 3.1]{MG1}. \n\n\\begin{lem}\\label{l2}\nFor $j\\ge 2$, the rows of Table 2 for which $L_{k+j}$ is the largest summand are those numbered from $F_{j}+1$ to $F_{j+1}$ inclusive. \n\\end{lem}\n\n\\begin{lem}\nFor $j\\ge 1$, we have\n$$q_k(j+1)-q_k(j) \\ =\\ \\begin{cases}L_{k+1}, & \\mbox{ if }A \\mbox{ is the }j\\mbox{th character of }S,\\\\\nL_{k+2}, &\\mbox{ if }B \\mbox{ is the }j\\mbox{th character of }S.\\end{cases}$$\n\\end{lem}\n\n\\begin{proof}\nWe prove by induction. \\textit{Base cases:} it is easy to check that the statement of the lemma is true for $j$ such that $1\\le j\\le F_4-1$. \\textit{Inductive hypothesis:} assume that it is true for $1\\le j\\le F_m-1$ for some $m\\ge 4$. From Lemma \\ref{l2}, the first $F_{m-1}$ rows of Table 2 are those for which the largest summand is no greater than $L_{k+m-2}$. Also, the rows for which $L_{k+m}$ is the largest summand are those numbered from $F_m+1$ to $F_{m+1}$ inclusive. Therefore, the ordering of the rows implies that \n$q_k(i+F_m) \\ =\\ q_k(i) + L_{k+m}$,\nfor $i = 1, 2, \\ldots, F_{m-1}$. Hence, for $i = 1, 2, \\ldots, F_{m-1}-1$, we have\n\\begin{align*}\n    q_k(i+1+F_m) - q_k(i+F_m) &\\ =\\ (q_k(i+1)+L_{k+m}) - (q_k(i) + L_{k+m})\\ =\\ q_k(i+1) - q_k(i).\n\\end{align*}\nBy the construction of $S$, the substring comprising its first $F_{m-1}$ characters is identical to the substring of its characters numbered from $F_m+1$ to $F_{m+1}$ inclusive. Thus, the lemma is true for $F_m+1\\le j\\le F_{m+1}-1$. It remains to show that the lemma is true for $j = F_m$. We have\n\\begin{align*}\n    q_k(F_m+1) - q_k(F_m)\\ =\\begin{cases}\\ L_{k+m} - (L_{k+m-1}+L_{k+m-3} + \\cdots + L_{k+3}) \\ =\\ L_{k+2},&\\mbox{ if }m\\mbox{ is even,}\\\\\n   L_{k+m} - (L_{k+m-1}+L_{k+m-3} + \\cdots + L_{k+2})\\ =\\ L_{k+1},&\\mbox{ if }m\\mbox{ is odd.}\\end{cases}\n\\end{align*}\nBy Remark \\ref{r1} item (1), we know that the lemma is true for $j = F_{m}$, completing the proof.  \n\\end{proof}\n\nWe are ready to prove Theorem \\ref{m2}.\n\n\\begin{proof}[Proof of Theorem \\ref{m2}] We consider three cases. \n\\mbox{ }\\newline\n\nCase 1: $k = 0$. By Lemma \\ref{k1}, we have $\\mathcal{X}_0 \\ =\\ \\{2+a(n)L_2 + b(n)L_3: n\\ge 0\\}$, where $a(n)$ and $b(n)$ denote the number of $A$'s and $B$'s, respectively, amongst the first $n$ characters in the golden string. Using Remark \\ref{r1} item (2), we have\n\\begin{align*}\n    \\mathcal{X}_0 &\\ =\\ \\left\\{2+3\\left(n-\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\right)+4\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}\\ =\\ \\left\\{2+3n+\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\,:\\, n\\ge 0\\right\\}.\n\\end{align*}\nIt is clear that $Z(0) = \\mathcal{X}_0$; hence, the statement of the lemma is true when $k = 0$. \n\\mbox{ }\\newline\n\nCase 2: $k = 1$. Using a similar reasoning as above, we have\n\\begin{align*}\n    \\mathcal{X}_1 &\\ =\\ \\left\\{1+L_2\\left(n-\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\right)+L_3\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}\\\\\n    &\\ =\\ \\left\\{1+3\\left(n-\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\right)+4\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}\\ =\\ \\left\\{3n+\\left\\lfloor \\frac{n+\\Phi^2}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}.\n\\end{align*}\nIt is clear that $Z(1) = \\mathcal{X}_1$; hence, the statement of the lemma is true when $k = 1$. \n\\mbox{ }\\newline\n\nCase 3: $k\\ge 2$. Using a similar reasoning as above, we have\n\\begin{align*}\n    \\mathcal{X}_k &\\ =\\ \\left\\{L_k+L_{k+1}\\left(n-\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\right)+L_{k+2}\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\, :\\, n\\ge 0\\right\\}\\\\\n    &\\ =\\ \\left\\{L_k\\left(1+\\left\\lfloor \\frac{n+1}{\\Phi}\\right\\rfloor\\right)+nL_{k+1}\\, :\\, n\\ge 0\\right\\}\\ =\\ \\left\\{L_k\\left\\lfloor \\frac{n+\\Phi^2}{\\Phi}\\right\\rfloor+nL_{k+1}\\, :\\, n\\ge 0\\right\\}.\n\\end{align*}\nIf $k\\ge 3$, the numbers in $\\{L_0, L_1, \\ldots, L_{k-2}\\}$ are used to obtain the partitions of all integers for which the largest summand is no greater than $L_{k-2}$. In particular, such partitions generate all integers from $1$ to $L_{k-1}-1$ inclusive. Furthermore, such partitions can be appended to any partition having $L_k$ as its smallest summand to produce another partition. Therefore,\n\\begin{align*}\n    Z(k) \\ =\\ \\left\\{L_k\\left\\lfloor \\frac{n+\\Phi^2}{\\Phi}\\right\\rfloor+nL_{k+1} + j\\, :\\, n\\ge 0\\mbox{ and } 0\\le j\\le L_{k-1}-1\\right\\},\n\\end{align*}\nas desired. It is easy to check that this formula is also true for $k = 2$. \n\\end{proof}\n\n\\section{Proportion of Nonunique Partitions}\nLet $c(N)$ count the number of numbers that are not uniquely represented in the Lucas sequence and are at most $N$. We want to show that $\\displaystyle\\lim_{N\\rightarrow \\infty}\\frac{c(N)}{N} = \\frac{1}{1+3\\Phi}$, where $\\Phi = (1+\\sqrt{5})/2$ is the golden ratio. Note that \\cite[Lemma 3]{B2} says we can make the Lucas partition unique by requiring that not both $L_0$ and $L_2$ appear in the partition. Therefore, if a number has two partitions, then one of the partition starts with $L_0+L_2$. If we can characterize all of these numbers and find a formula for $c(N)$ in terms of $N$, we are done. Call the set of these numbers $K$. We form the following table listing all of such numbers in increasing order. Let $q_k(j)$ be the $jth$ smallest number in $K$. \n\n\\begin{center}\n\\begin{tabular}{cccccccccccc}\n Row  &               &          &     &        &    &         &     &        &  &          &    \\\\\n \\hline\n1     &               & $L_0+L_2$    &     &        &    &         &     &        &  &         &     \\\\\n2     &               & $L_0+L_2$    &     & $L_4$  &    &         &     &        &  &         &    \\\\\n3     &               & $L_0+L_2$    &     &        &    & $L_5$   &     &        &  &         &     \\\\\n$4$   &               & $L_0+L_2$    &     &        &    &         &     & $L_6$  &  &         &     \\\\\n$5$   &               & $L_0+L_2$    &     & $L_4$  &    &         &     & $L_6$  &  &         &       \\\\\n$6$   &               & $L_0+L_2$    &     &        &    &         &     &        &  &  $L_7$  &   \\\\\n$7$   &               & $L_0+L_2$    &     & $L_4$  &    &         &     &        &  &  $L_7$  &     \\\\\n$8$   &               & $L_0+L_2$    &     &        &    & $L_5$   &     &        &  &  $L_7$  &    \n\\end{tabular}\n\\end{center}\n\n\\begin{center}\nTable 3. The partitions of the positive integers having $L_0$ and $L_2$ as their smallest summands.  \n\\end{center}\n\nObserve that Table 3 has the same structure as Table 1. Therefore, Lemma \\ref{k1} applies with a change of index. In particular, we have the following.  \n\n\n\n\\begin{lem}\\label{k2}\nFor $j\\ge 1$, we have\n$$q_k(j+1)-q_k(j) \\ =\\ \\begin{cases}L_{3}, & \\mbox{ if }A \\mbox{ is the }j\\mbox{th character of }S,\\\\\nL_{4}, &\\mbox{ if }B \\mbox{ is the }j\\mbox{th character of }S.\\end{cases}$$\n\\end{lem}\nTherefore, we can write \n$$K \\ =\\ \\{L_0+L_2 + a(n)L_3 + b(n)L_4: n\\ge 0\\},$$\nwhere $a(n)$ and $b(n)$ denote the number of $A$'s and $B$'s, respectively, amongst the first $n$ characters in the golden string. Hence, \n\\begin{align*}K &\\ =\\ \\{5 + 4(n-\\floor{(n+1)/\\Phi})+7\\floor{(n+1)/\\Phi}\\,:\\, n\\ge 0\\}\\ =\\ \\{5+4n+3\\floor{(n+1)/\\Phi}\\,:\\, n\\ge 0\\}.\\end{align*}\n\nNow, we are ready to compute the limit. \n\n\\begin{proof}[Proof of Theorem \\ref{m3}]The number of integers with two partitions up to a number $N$ is exactly\n$\\#\\{n\\ge 0\\,|\\, 5+4n+3\\floor{(n+1)/\\Phi}\\le N\\}$.\nThe number is found to be $\\frac{N-1}{4+3/\\Phi}$ within an error of at most $1$. Therefore, as claimed, the limit is \n$$\\lim_{N\\rightarrow \\infty}\\frac{1}{N}\\frac{N-1}{4+3/\\Phi}\\ =\\ \\frac{1}{4+3/\\Phi} \\ =\\ \\frac{1}{1+3\\Phi}.$$\n\\end{proof}\n\nAmong the first $N$ natural numbers, we see how $\\alpha = \\frac{1}{3\\Phi + 1} \\approx 0.17082$ estimates the proportion of natural numbers within this range that do not have unique non-consecutive partitions in the Lucas sequence.~The data we collect is shown in Table 4.\n\\begin{center}\n    \\begin{tabular}{|c|c|c|}\n    \\hline\n      N & $c\\left(N \\right)$ & $\\beta\\left(N\\right)$ \\\\\n      \\hline\n      10 & 1 & 10.000 \\% \\\\ \\hline\n      100 & 17 & 17.000\\% \\\\ \\hline\n      1,000 & 171 & 17.100\\% \\\\ \\hline\n      10,000 & 1,708 & 17.080\\% \\\\ \\hline\n      $10^5$ & 17,082 & 17.082\\%\\\\ \\hline\n      $10^6$ & 170,820 & 17.082\\%\\\\ \\hline\n    \\end{tabular}\n\\end{center}\n\n\\begin{center}\nTable 4. Proportion $\\beta\\left(N\\right)$ of the first $N$ natural numbers that do not have unique non-consecutive partitions in the Lucas sequence.\n\\end{center}\n\n\n\n", "meta": {"timestamp": "2021-02-24T02:06:57", "yymm": "2004", "arxiv_id": "2004.08316", "language": "en", "url": "https://arxiv.org/abs/2004.08316", "abstract": "Zeckendorf proved that every positive integer has a unique partition as a sum of non-consecutive Fibonacci numbers. Similarly, every natural number can be partitioned into a sum of non-consecutive terms of the Lucas sequence, although such partitions need not be unique. In this paper, we prove that a natural number can have at most two distinct non-consecutive partitions in the Lucas sequence, find all positive integers with a fixed term in their partition, and calculate the limiting value of the proportion of natural numbers that are not uniquely partitioned into the sum of non-consecutive terms in the Lucas sequence.", "subjects": "Number Theory (math.NT)", "title": "On Zeckendorf Related Partitions Using the Lucas Sequence", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9895109090493823, "lm_q2_score": 0.8558511469672594, "lm_q1q2_score": 0.8468740464465293}}
{"url": "https://arxiv.org/abs/2110.01132", "title": "Extrema of Luroth Digits and a zeta function limit relation", "abstract": "We describe how certain properties of the extrema of the digits of Luroth expansions lead to a probabilistic proof of a limiting relation involving the Riemann zeta function and the Bernoulli triangles. We also discuss trimmed sums of Luroth digits. Our goal is to show how direct computations in this case lead to explicit formulas and some interesting discussions of special functions.", "text": "\\section{\\@startsection {section}{1}{\\z@}\n{-30pt \\@plus -1ex \\@minus -.2ex}\n{2.3ex \\@plus.2ex}\n{\\normalfont\\normalsize\\bfseries\\boldmath}}\n\n\\renewcommand\\subsection{\\@startsection{subsection}{2}{\\z@}\n{-3.25ex\\@plus -1ex \\@minus -.2ex}\n{1.5ex \\@plus .2ex}\n{\\normalfont\\normalsize\\bfseries\\boldmath}}\n\n\\renewcommand{\\@seccntformat}[1]{\\csname the#1\\endcsname. }\n\n\\makeatother\n\n\\newtheorem{theorem}{Theorem}\n\\newtheorem{lemma}{Lemma}\n\\newtheorem{proposition}{Proposition}\n\\newtheorem{corollary}{Corollary}\n\n\\theoremstyle{definition}\n\\newtheorem{definition}{Definition}\n\\newtheorem{conjecture}{Conjecture}\n\\newtheorem{remark}{Remark}\n\n\n\\begin{document}\n\n\n\n\\begin{center}\n\\uppercase{\\bf Extrema of Luroth Digits and a zeta function limit relations}\n\\vskip 20pt\n{\\bf Jayadev S. Athreya\\footnote{Supported by National Science Foundation Grant DMS 2003528}}\\\\\n{\\smallit Department of Mathematics and Department of Comparative History of Ideas, University of Washington, Seattle, WA, USA.}\\\\\n{\\tt jathreya@uw.edu}\\\\ \n{\\bf Krishna B. Athreya}\\\\\n{\\smallit Department of Mathematics and Department of Statistics, Iowa State University, Ames, IA, USA.}\\\\\n{\\tt kbathreya@gmail.com}\\\\ \n\\end{center}\n\\vskip 20pt\n\\centerline{\\smallit Received: , Revised: , Accepted: , Published: }\n\\vskip 30pt\n\n\n\\centerline{\\bf Abstract}\n\\noindent\nWe describe how certain properties of the extrema of the digits of Luroth expansions lead to a probabilistic proof of a limiting relation involving the Riemann zeta function and the Bernoulli triangles. We also discuss trimmed sums of Luroth digits. Our goal is to show how direct computations in this case lead to formulas and some interesting discussions of special functions. \n\n\\pagestyle{myheadings}\n\\markright{\\smalltt INTEGERS: 21 (2021)\\hfill}\n\\thispagestyle{empty}\n\\baselineskip=12.875pt\n\\vskip 30pt\n\n\\section{Introduction} \n\nLet $X_1, X_2, \\ldots X_k, \\ldots$ be a sequence of independent, identically distributed (IID) random variables taking values in the positive integers $\\mathbb N =\\{1, 2, 3, \\ldots \\}$ with $$P(X_1 = n) = \\frac{1}{n(n+1)}, n \\geq 1.$$ We call this distribution the \\emph{Luroth} distribution. The \\emph{Luroth map} $L: (0, 1] \\rightarrow (0, 1]$ is the piecewise-linear map given by $$L(x) = N(x) ( (N(x)+1) x - 1),$$ where $$N(x) =\\left \\lfloor \\frac 1 x \\right \\rfloor.$$ It is a nice exercise to check that $L$ preserves Lebesgue measure $m$, and that if $U$ is a uniform $(0, 1]$ random variable, the sequence $X_k = N(L^{k-1} (U))$ is a sequence of IID random variables with the above distribution, since $$P(X_1 = n) = m\\left(N^{-1}(n)\\right) = m\\left(\\frac{1}{n+1}, \\frac{1}{n}\\right] = \\frac{1}{n(n+1)}.$$ \n\nThis map and its associated digit expansion was introduced by Luroth in~\\cite{Luroth}.  Subsequently, Luroth random variables have been extensively studied,  since they provide an attractive motivating example at the intersection of many questions on number theory, probability theory, and dynamical systems and ergodic theory. The Luroth map is in some sense a kind of linearized version of the Gauss map (see Section~\\ref{sec:Gauss}), where, as we remark above, the resulting digit expansion of numbers have the appealing property that the digits are \\emph{independent} random variables. \n\nThis makes it often possible to do very explicit computations with Luroth random variables. See~\\cite{JDv} for an introduction to the ergodic properties of the map $L$, ~\\cite{Galambos} for a discussion of different types of digit expansions, and, for example~\\cite{Giuliano} and~\\cite{Tan} and the references within for more recent work on Luroth and related expansions from the probabilistic and dynamical perspectives respectively.\n\nOur results center on understanding the limiting behavior of \\emph{maxima} of sequences of IID Luroth random variables, and the convergence behavior of appropriately centered and scaled sums. These are classical topics in the theory of sequences of random variables, and focusing on this setting leads to some interesting computations.\n\nUsing the fact that the Luroth random variables are \\emph{heavy-tailed}, we will show that the probability $\\rho_k$ that the maximum $M_k$ of the first $k$ elements $\\{X_1, \\ldots, X_k\\}$ of a a sequence of Luroth variables is \\emph{unique} tends to $1$ as $k \\rightarrow \\infty.$ Our main new result, Theorem~\\ref{theorem:rhok}, shows how to \\emph{explicitly} compute this probability $\\rho_k$, giving an interesting relationship between this probability, the Riemann zeta function, and partial sums of binomial coefficients which are entries in what is known as Bernoulli's triangle. Using the fact that $\\rho_k \\rightarrow 1$, we obtain a new limiting relation, Corollary~\\ref{cor:zeta}, between these coefficients and the Riemann zeta function. \n\n\nOur other main result considers the behavior of the sample sums $$S_k = \\sum_{i=1}^k X_i.$$ Since $E(X_1) = \\infty$, the law of large numbers shows that $S_k/k$ tends to infinity with probability one. Inspired by work of Diamond-Vaaler, who studied the corresponding \\emph{trimmed sum} for continued fraction expansions, and using a result of Mori~\\cite{Mori} we show in Theorem~\\ref{theorem:trimmed} that by removing the maximum, and normalizing by $k\\log k$, we have, with probability $1$, as $k \\rightarrow \\infty$,  $$\\frac{S_k-M_k}{k\\log k} \\rightarrow 1.$$ Note that the asymptotic uniqueness result says that asymptotically, we are removing the \\emph{unique} largest summand in $S_k$ by subtracting $M_k$. Let $$M_k = \\max(X_1, \\ldots, X_k)$$ be the sample maximum, and let $$\\rho_k = P(\\mbox{ there exists } !  1 \\le i \\le k \\mbox{ such that }X_i = M_k)$$ be the probability that this maximum is achieved exactly once.  Our first main observation is the following explicit formula for $\\rho_k$. \n\\begin{theorem}\\label{theorem:rhok} For $k \\geq 1$, \\begin{equation}\\label{eq:rhok} \\rho_k = k\\left (2^{k-1} + \\sum_{j=2}^k T(k-1, k-j) \\zeta(j) (-1)^{j+1}\\right),\\end{equation}  where $\\zeta(j) = \\sum_{n=1}^{\\infty} n^{-j}, j > 1$ is the Riemann zeta function and for $l \\geq j$ positive integers, $$T(l, j) = \\sum_{i=0}^j \\binom{l}{i}.$$\n\\end{theorem}\n\nBy \\cite{AF, ESS}, we have $$\\lim_{k \\rightarrow \\infty} \\rho_k = 1.$$ Thus, as a corollary, we obtain the following limiting relation involving the coefficients $T(l, j)$ and values of the Riemann zeta function. \n\n\\begin{corollary}\\label{cor:zeta} We have\n\\begin{equation}\\label{eq:zeta}\\lim_{k \\rightarrow \\infty}  k\\left (2^{k-1} + \\sum_{j=2}^k T(k-1, k-j) \\zeta(j) (-1)^{j+1}\\right) = 1.\\end{equation}\n\n\\end{corollary}\n\nThe coefficients $T(l, j)$, partial sums of binomial coefficients, are entries in what is known as \\emph{Bernoulli's triangle}; see~\\cite{OEIS-A008949}. Note that $T(l, 0) = 1$, $T(l, l) = 2^l$, and $T(l, l-1) = 2^{l} -1$. Let $$S_k = \\sum_{i=1}^k X_i$$ be the sample sum. Since $E(X_1) = \\infty$, we have, by the strong law of large numbers, $$\\lim_{k \\rightarrow \\infty} \\frac{S_k}{k} = +\\infty.$$ However, if we subtract the sample maximum, we have the following almost sure result.\n\n\\begin{theorem}\\label{theorem:trimmed} With probability $1$, $$\\lim_{k \\rightarrow \\infty}  \\frac{S_k - M_k}{k \\log k} = 1.$$\n\n\\end{theorem}\n\n\\noindent This is a consequence of a result of Mori~\\cite{Mori}. The proof features a nice appearance of the Lambert $W$-function. There is a (more difficult) analogous result~\\cite{DV} for continued fraction digits, where the limiting value is $\\log 2$. There are substantial additional difficulties in that setting since continued fraction digits of a randomly chosen $x \\in (0, 1)$ do not form an IID sequence. \n\n\n\n\\section{Uniqueness of Maxima}\\label{sec:maxproof}\n Suppose $Y_1, \\ldots, Y_k, \\ldots$ is a sequence of IID $\\mathbb N$-valued random variables with $$P(Y_1 = n) = p_n.$$ Let $$\\tau_n = \\sum_{m \\geq n} p_m = P(Y_1 \\geq n).$$ Let $\\rho_{k, m}$ denote the probability that the sample maximum $M_k = \\max(Y_1, \\ldots, Y_k) = m$ and is achieved exactly once. Then $$\\rho_{k,m} = k p_m (1-\\tau_m)^{k-1}.$$ Note that $\\rho_k$, the probability that the sample maximum $M_k$ is achieved exactly once by $Y_1, \\ldots, Y_k$, is then given by $$\\rho_k = \\sum_{m=1}^{\\infty} \\rho_{k,m}.$$ Also note that (although it is not a probability) we have $$\\sum_{k=1}^{\\infty} \\rho_{k,m} = p_m \\sum_{k=1}^{\\infty} k(1-\\tau_m)^{k-1} = \\frac{p_m}{\\tau_m^2}.$$ Specializing to Luroth sequences, we have $p_m = \\frac{1}{m(m+1)}$ and $\\tau_m = \\frac{1}{m}$. Thus $$\\rho_{k,m} = \\frac{k}{m(m+1)} \\left( 1 - \\frac 1 m \\right)^{k-1} = k \\frac{(m-1)^{k-1}}{m^k(m+1)} .$$ Let $$Q_k(m) = \\frac{1}{k} \\rho_{k, m} = \\frac{(m-1)^{k-1}}{m^k(m+1)}.$$ Th next lemma shows how to expand $Q_k$ as a partial fraction.\n\n\\begin{lemma}\\label{lem:partial} We have \\begin{equation}\\label{eq:qk}Q_k(m) = 2^{k-1} \\left( \\frac 1 m - \\frac{1}{m+1} \\right) + \\sum_{j=2}^k T(k-1, k-j) (-1)^{j+1} \\frac{1}{m^j}.\\end{equation}\n\\end{lemma}\n\n\\begin{proof} Note that $$T(k-1, 0) = 1, T(k-1, k-1) = 2^{k-1}, T(k-1, k-2) = 2^{k-1} - 1$$ and $$Q_{k+1}(m) = \\left( 1 - \\frac 1 m\\right) Q_{k}(m).$$ Our proof proceeds by induction. For $k=1$, we have $$Q_1(m) = \\rho_{1, m} = \\frac{1}{m(m+1)}.$$ The right hand side of  Equation (\\ref{eq:qk}) for $k=1$ is $$2^{1-1} \\left (\\frac 1 m - \\frac{1}{m+1} \\right) = \\frac{1}{m(m+1)},$$ since the index of the sum starts at $j=2$, making the sum empty. Thus the base case ($k=1$) is verified. We now assume that Equation (\\ref{eq:qk}) is true for $k$, and we need to prove it for $k+1$. Let $$c(j, k) = (-1)^{j+1} T(k-1, k-j).$$ Note that\n\n\\begin{align}\\label{eq:induction} Q_{k+1}(m) &= \\left( 1 - \\frac 1 m\\right) Q_{k}(m)  \\nonumber\n= \\left( 1 - \\frac 1 m\\right) \\left(2^{k-1} \\left( \\frac 1 m - \\frac{1}{m+1} \\right) + \\sum_{j=2}^k c(j,k) \\frac{1}{m^j}\\right) \\\\ \\nonumber\n&= 2^{k-1}\\frac{1}{m(m+1)} - \\frac{2^{k-1}}{m}\\left( \\frac 1 m - \\frac{1}{m+1}\\right) + \\sum_{j=2}^{k}\\left( \\frac{c(j,k) }{m^j} -\\frac{c(j-1, k) }{m^{j+1}}\\right)\\\\ \\nonumber \n&= 2^k \\frac{1}{m(m+1)} - \\frac{(2^{k-1} - c(2, k))}{m^2} + \\sum_{j=3}^{k+1} \\frac{\\left(c(j, k)- c(j-1, k) \\right)}{m^j} . \\\\\n\\end{align}\n\n\\noindent Now note that for $j \\geq 2$, $$c(j, k) = c(j, k-1) - c(j-1, k-1)$$ and $$c(k+1, k) = 0, c(2, k) = 1-2^{k-1}, c(k,k) = (-1)^{k+1}, c(1, k) = 2^{k-1}.$$ Plugging these into Equation (\\ref{eq:induction}), we have our result.\n\\end{proof}\n\n\\subsection{Proof of Theorem~\\ref{theorem:rhok}} To prove Theorem~\\ref{theorem:rhok}, we note that $$\\sum_{m=1}^{\\infty} \\frac{1}{m(m+1)} = 1;  \\sum_{m=1}^{\\infty} \\frac{1}{m^j} = \\zeta(j), j \\geq 2.$$ Using  Lemma~\\ref{lem:partial}, and summing over $m$, we obtain Equation (\\ref{eq:rhok}). To obtain Corollary~\\ref{cor:zeta}, it was shown in~\\cite{AF, ESS} that $$\\rho_k \\rightarrow 1 \\mbox{ if and only if } p_m/\\tau_m \\rightarrow 0.$$ In our setting, $\\rho_m/\\tau_m = \\frac{1}{m+1}$, so the condition is satisfied, and we obtain Corollary~\\ref{cor:zeta}. In Figure~\\ref{fig:rhok}, we show numerical values of $\\rho_k$ for $2 \\le k \\le 40$ (note that $\\rho_1 = 1$).\n\n\\begin{figure}[h!] \\includegraphics[width=0.9\\textwidth]{rhokgraph.pdf}\n\\caption{$\\rho_k$ for $k =2$ to $k=40$. \\medskip}\\label{fig:rhok}\n\\end{figure}\n\n\n\\section{Trimmed sums}\\label{sec:trimproof}\nIn this section, we prove Theorem~\\ref{theorem:trimmed}. We first note that using standard limit theorems in probability theory, we can show that $$\\frac{S_k - M_k}{k \\log k} \\xrightarrow{d} 1.$$ The random variables $X_n$ are in the domain of attraction of a \\emph{stable law} of order $1$. Using~\\cite[Theorem 11.2.3]{AL}, we have $$\\frac{S_k - k\\log k}{k} \\xrightarrow{d} C,$$ where $C$ is a Cauchy random variable. Thus, $$\\frac{S_k}{k \\log k}  = \\left(\\frac{1}{\\log k}\\frac{(S_k - k \\log k)}{k} + 1 \\right) \\xrightarrow{d} 1.$$\n\n\nNext, note that for $c >0$, $$\\lim_{k \\rightarrow \\infty} P\\left(\\frac{M_k}{k} < c\\right) = \\lim_{k \\rightarrow \\infty} \\left( 1 - \\frac{1}{ck}\\right)^{k} = e^{-1/c}.$$ Therefore $$\\frac{M_k}{k\\log k} \\xrightarrow{d} 0.$$\n\nPutting these together, we obtain $$\\frac{S_k - M_k}{k \\log k} \\xrightarrow{d} 1,$$ and therefore we also have this convergence in probability.\n\nTo replace convergence in probability with almost sure convergence, we use a result of Mori~\\cite[Theorem 1]{Mori}, with, in his notation, $$r=1, A(x) = x \\log x.$$ Mori studies the behavior of the sample sums with the first $r$ largest terms removed, normalized by $A(n)$, where $A$ is an absolutely continuous increasing function such that there is an $0< \\alpha< 2$, with $$A(x)x^{-\\frac{1}{\\alpha}} \\mbox{ increasing and } \\sup \\frac{A(2x)}{A(x)} < \\infty.$$ Keeping our notation consistent with ~\\cite[Section 1]{Mori}, we write $A(x) = x \\log x$, and note that this is a non-decreasing absolutely continuous function, with inverse function $B(x) = \\frac{x}{W(x)}$, where $W(x)$ is the Lambert $W$ function, that is, the inverse of the function $we^w$. Putting $$\\mathcal F(x) = P(X_1 > x) = \\frac{1}{\\lfloor x \\rfloor + 1}$$ and $$J_2 = \\int_{0}^{\\infty} \\mathcal F^2(x) dB^2(x) = \\sum_{n=2}^{N} \\frac{1}{n^2} \\left( \\frac{n^2}{W(n)^2} - \\frac{(n-1)^2}{W(n-1)^2}\\right),$$ Mori's result shows that if $J_2 < \\infty$, that almost surely $$\\frac{S_k-M_k}{A(k)} - c_k  \\rightarrow 0,$$ where $$c_k = \\frac{k}{A(k)} E(X_1| X_1 < A(k)).$$ In our setting, $$c_k \\sim \\frac{k}{k \\log k} \\log(k \\log k) \\rightarrow 1 \\mbox{ as } k \\rightarrow \\infty.$$ So if we can show $J_2 < \\infty$, we have completed our proof of Theorem~\\ref{theorem:trimmed}. From~\\cite{lambertw}, we see that for $x >>1$, $$W(x) = \\log x - \\log \\log x  + o(1).$$ Note that $$dB^2(x) = 2 B(x) B'(x) dx = \\frac{x}{W(x) (W(x) + 1)} dx.$$ Thus we have $$J_2 \\le \\int_1^{\\infty} \\frac{2}{x^2} B(x) B'(x) dx \\sim 2 \\int_{1}^{\\infty} \\frac{1}{x W(x) (W(x) + 1)}  =  2 \\left(\\frac{1}{W(1)}\\right).$$ In particular,  $J_2 < \\infty$. In Figure~\\ref{fig:lambert}, we show the partial sums for $J_2$.\n\n\\begin{figure}[h!]\\includegraphics[width=0.9\\textwidth]{LambertMoriSum.pdf}\\caption{Partial sums of $\\sum_{n=2}^{N} \\frac{1}{n^2} \\left( \\frac{n^2}{W(n)^2} - \\frac{(n-1)^2}{W(n-1)^2}\\right)$, $N=3, \\ldots, 1000$. \\medskip}\\label{fig:lambert}\n\n\\end{figure}\n\n\\section{Continued Fractions}\\label{sec:Gauss} \\noindent Similar questions can be asked for the \\emph{Gauss map} $G: (0, 1) \\rightarrow (0, 1)$, $$G(x) = \\left\\{ \\frac 1 x \\right\\},$$ and the associated continued fraction expansion of a number $x \\in [0, 1)$. Recall that for $x \\in (0, 1)$, we can set, for $n \\geq 1$, $$a_n = a_n(x) = \\left \\lfloor \\frac{1}{G^{n-1}(x)} \\right \\rfloor,$$ and we can write $$x = \\frac{1}{a_1 + \\frac{1}{a_2 + \\frac{1}{a_3 + \\ldots}}}.$$ The probability measure $\\mu_G$ given by $$d\\mu_G(x) = \\frac{1}{\\log 2} \\frac{dx}{1+x}$$ is an absolutely continuous ergodic invariant measure for $G$, and as noted after Theorem~\\ref{theorem:trimmed}, Diamond and Vaaler~\\cite{DV} proved a similar result for the digits $a_n$ for an $x$ chosen at random according to $\\mu_G$. This is more difficult than our setting since although the digits form a stationary sequence, they are no longer IID. Regarding the uniqueness of sample maxima, if we define $$\\rho_k = \\mu_G\\left(x \\in (0, 1): \\mbox{ there exists } ! 1 \\le i \\le k \\mbox{ such that } a_i (x)= \\max_{1 \\le j \\le k} a_j(x) \\right),$$ we conjecture that $$\\lim_{k \\rightarrow \\infty} \\rho_k = 1.$$ We believe this conjecture since although the sequence $\\{a_i\\}$ is not IID, it is rapidly mixing, and the distribution of $a_i$ is heavy-tailed. The Gauss map is associated in a natural way to geodesic flow on the modular surface, and we believe similar results should also hold for other continued fraction type expansions associated to geodesic flow on other hyperbolic surfaces, for example, the Rosen continued fractions.\n\n\\bigskip\n\n\n \\noindent {\\bf Acknowledgements.} We thank Bruce Berndt, Alex Kontorovich, and Jeffrey Lagarias for useful discussions. Lagarias pointed out to us some interesting asymptotics involving binomial coefficients and zeta functions in~\\cite[Theorem 2]{BL}. We thank Avanti Athreya, the anonyomous referee, and the Integers Editorial Office for suggestions that greatly improved the exposition of the paper. The first author thanks the National Science Foundation for its support via grant NSF DMS 2003528. The first author is a faculty member at the University of Washington, which is on the lands of the Coast Salish people, and touches the shared waters of all tribes and bands within the Suquamish, Tulalip and Muckleshoot nations.\n\n", "meta": {"timestamp": "2021-10-05T02:24:56", "yymm": "2110", "arxiv_id": "2110.01132", "language": "en", "url": "https://arxiv.org/abs/2110.01132", "abstract": "We describe how certain properties of the extrema of the digits of Luroth expansions lead to a probabilistic proof of a limiting relation involving the Riemann zeta function and the Bernoulli triangles. We also discuss trimmed sums of Luroth digits. Our goal is to show how direct computations in this case lead to explicit formulas and some interesting discussions of special functions.", "subjects": "Probability (math.PR); Dynamical Systems (math.DS); Number Theory (math.NT)", "title": "Extrema of Luroth Digits and a zeta function limit relation", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9828232884721166, "lm_q2_score": 0.8615382076534742, "lm_q1q2_score": 0.8467398143903608}}
{"url": "https://arxiv.org/abs/1907.13292", "title": "The character graph of a finite group is perfect", "abstract": "For a finite group $G$, let $\\Delta(G)$ denote the character graph built on the set of degrees of the irreducible complex characters of $G$. In graph theory, a perfect graph is a graph $\\Gamma$ in which the chromatic number of every induced subgraph $\\Delta$ of $\\Gamma$ equals the clique number of $\\Delta$. In this paper, we show that the character graph $\\Delta(G)$ of a finite group $G$ is always a perfect graph. We also prove that the chromatic number of the complement of $\\Delta(G)$ is at most three.", "text": "\\section{Introduction}\n$\\noindent$ Let $G$ be a finite group . Also let ${\\rm cd}(G)$ be the set of all character degrees of $G$, that is,\n ${\\rm cd}(G)=\\{\\chi(1)|\\;\\chi \\in {\\rm Irr}(G)\\} $, where ${\\rm Irr}(G)$ is the set of all complex irreducible characters of $G$. The set of prime divisors of character degrees of $G$ is denoted by $\\rho(G)$. It is well known that the\n character degree set ${\\rm cd}(G)$ may be used to provide information on the structure of the group $G$. For example, Ito-Michler's Theorem \\cite{[C]} states that if a prime $p$ divides no character degree of a finite group $G$, then $G$ has a\n normal abelian Sylow $p$-subgroup. Another result due to Thompson \\cite{[D]} says that if a prime $p$ divides\n every non-linear character degree of a group $G$, then $G$ has a normal $p$-complement.\n\nA useful way to study the character degree set of a finite group $G$ is to associate a graph to ${\\rm cd}(G)$.\nOne of these graphs is the character graph $\\Delta(G)$ of $G$ \\cite{[I]}. Its vertex set is $\\rho(G)$ and two vertices $p$ and $q$ are joined by an edge if the product $pq$ divides some character degree of $G$. We refer the readers to a survey by Lewis \\cite{[M]} for results concerning this graph and related topics.\n\nLet $\\Gamma=(V(\\Gamma),E(\\Gamma))$ be a finite simple graph with the vertex set $V(\\Gamma)$ and edge set $E(\\Gamma)$.  A clique of $ \\Gamma $ is a set of mutually adjacent vertices, and that the maximum size of a clique of $ \\Gamma $, the clique number of $ \\Gamma $, is denoted by $ \\omega(\\Gamma) $. Minimum number of colors needed to color vertices of the graph $ \\Gamma $ so that any two adjacent vertices of $ \\Gamma $ have different colors, is called the chromatic number of $ \\Gamma $ and denoted by $ \\chi(\\Gamma) $. Clearly $\\omega(\\Gamma)\\leqslant \\chi(\\Gamma)$ for any graph $\\Gamma$. The graph $\\Gamma$ is perfect if $\\omega(\\Delta)=\\chi(\\Delta)$, for every induced subgraph $\\Delta$ of $\\Gamma$.\n\nThe theory of perfect graphs relates the concept of graph colorings to the concept of cliques. Aside from having an interesting structure, perfect graphs are considered important for three reasons. First, several common classes of graphs are known to always be perfect. For instance, bipartite graphs, chordal graphs and comparability graphs are perfect. Second, a number of important algorithms only work on perfect graphs. Finally, perfect graphs can be used in a wide variety  of applications, ranging from scheduling to order theory to communication theory.\n\nOne of the turning point for the investigation on the character degree graph is the \"Three-Vertex Theorem\" by Palfy \\cite{palfy}: the complement of $\\Delta(G)$ does not contain any triangle whenever $G$ is a finite solvable group. In a recent paper  \\cite{1} this was extended by showing that, under the same solvability assumption, the complement of $\\Delta(G)$ dose not contain any cycle of odd length, which is equivalent to say that the complement of $\\Delta(G)$ is a bipartite graph. Thus as the complement of a perfect graph is perfect \\cite{lovasz}, the character graph $\\Delta(G)$ of a solvable group $G$ is perfect. This argument motivates an interesting problem: is the character graph of an arbitrary finite group perfect?. In this paper, we wish to solve this problem.\\\\\n\n\\noindent \\textbf{Theorem A.}  \\textit{Let $G$ be a finite group then $\\Delta(G)$ is a perfect graph.\n}\\\\\n\nTheorem A will be used to determine the chromatic number of the complement of $\\Delta(G)$.\\\\\n\n\\noindent \\textbf{Corollary B.}  \\textit{ Suppose $G$ is a finite group. Then the chromatic number of the complement of $\\Delta(G)$ is at most three.\n}\n\\section{Preliminaries}\n$\\noindent$ In this paper, all groups are assumed to be finite and all\ngraphs are simple and finite. For a finite group $G$, the set of prime divisors of $|G|$ is denoted by $\\pi(G)$.  Also note that for an integer $n\\geqslant 1$,  the set of prime divisors of $n$ is denoted by $\\pi(n)$. We begin with Corollary 11.29 of \\cite{[isa]}.\n\n\n\\begin{lemma}\\label{fraction}\nLet $ N \\lhd G$ and $\\chi \\in \\rm{Irr}(G)$. Let $\\varphi \\in \\rm{Irr}(N)$ be a constituent of $\\chi_N$. Then $\\chi(1)/\\varphi(1)$ divides $[G:N]$.\n\\end{lemma}\n\nLet $\\Gamma$ be a graph with vertex set $V(\\Gamma)$ and edge set\n$E(\\Gamma)$. A cycle on $n$ vertices $v_1, \\ldots,\nv_n$, $n \\geq 3$, is a graph whose vertices can be arranged in a\ncyclic sequence in such a way that two vertices are adjacent if\nthey are consecutive in the sequence, and are non-adjacent\notherwise. A cycle with $n$ vertices is said to be of length $n$\nand is denoted by $C_n$, i.e., $C_n : v_1, \\ldots, v_n, v_1$.\n The join $\\Gamma \\ast \\Delta$ of graphs $\\Gamma$\nand $\\Delta$ is the graph $\\Gamma \\cup \\Delta$ together with all\nedges joining $V (\\Gamma)$ and $V (\\Delta)$.  An independent set is a set of vertices of a graph\n$\\Gamma$ such that no two of them are adjacent. A maximum\nindependent set is an independent set of largest possible size.\nThis size is called the independence number of $\\Gamma$ and is\ndenoted by $\\alpha(\\Gamma)$. Finally we should mention that  the complement of $\\Gamma$ and the induced subgraph of $\\Gamma$ on $X\\subseteq V(\\Gamma)$\n are denoted by $\\Gamma^c$ and  $\\Gamma[X]$, respectively.  For more details, we\nrefer the reader to basic textbooks on the subject, for instance\n\\cite{BM}. Now we present some properties of perfect graphs.\n\n\\begin{lemma}\\label{comp}\\cite{lovasz}\nA graph $\\Gamma$ is perfect if and only if the complement of $\\Gamma$ is perfect.\n\\end{lemma}\n\n\\begin{lemma}\\label{hole}\\cite{[cr]}\nA graph $\\Gamma$ is perfect  if and only if it has no induced subgraph isomorphic either to a cycle of odd order at least 5, or to the complement of such a cycle.\n\\end{lemma}\n\nWe now state some relevant results on character graphs\nneeded in the next section.\n\n\\begin{lemma}\\label{pal}\nLet $G$ be a group and let $\\pi \\subseteq \\rho(G)$.\\\\\n\\textbf{a)} (Palfy's condition) \\cite {palfy} If $G$ is solvable and $|\\pi|\\geqslant 3$, then there exist two distinct primes $u, v$ in $\\pi$ and $\\chi \\in \\rm{Irr}(G)$ such that $uv | \\chi(1)$.\\\\\n\\textbf{b)} (Moreto-Tiep's condition) \\cite{moreto} If $|\\pi|\\geqslant 4$, then there exists $\\chi \\in \\rm{Irr}(G)$ such that $\\chi (1)$ is divisible by two distinct primes in $\\pi$.\n\\end{lemma}\n\n\\begin{lemma}\\label{chpsl}\\cite{[white]}\nLet $G\\cong \\rm{PSL}_2(q)$, where $q\\geqslant 4$ is a power of a prime $p$.\\\\\n\\textbf{a)}\n If $q$ is even, then $\\Delta(G)$ has three connected components, $\\{2\\}$, $\\pi(q-1)$ and $\\pi(q+1)$, and each component is a complete graph.\\\\\n\\textbf{b)}\n If $q>5$ is odd, then $\\Delta(G)$ has two connected components, $\\{p\\}$ and $\\pi((q-1)(q+1))$.\\\\\ni)\n The connected component $\\pi((q-1)(q+1))$ is a complete graph if and only if $q-1$ or $q+1$ is a power of $2$.\\\\\nii)\n  If neither of $q-1$ or $q+1$  is a power of $2$, then $\\pi((q-1)(q+1))$ can be partitioned as $\\{2\\}\\cup M \\cup P$, where $M=\\pi (q-1)-\\{2\\}$ and $P=\\pi(q+1)-\\{2\\}$ are both non-empty sets. The subgraph of $\\Delta(G)$ corresponding to each of the subsets $M$,$P$ is complete, all primes are adjacent to $2$, and no prime in $M$ is adjacent to any prime in $P$.\n \\end{lemma}\n\n\\begin{lemma}\\label{cycle} \\cite{AC}\nLet $G$ be a finite group, and let $\\pi$ be a subset of the vertex set of $\\Delta(G)$ such that $|\\pi|$ is an odd number larger than 1. Then $\\pi$ is the set of vertices of a cycle in $\\Delta(G)^c$ if and only if $O^{\\pi^\\prime}(G)=S\\times A$, where $A$ is abelian, $S\\cong \\rm{SL}_2(u^\\alpha)$ or $S\\cong \\rm{PSL}_2(u^\\alpha)$ for a prime $u\\in \\pi$ and a positive integer $\\alpha$, and the primes in $\\pi - \\{u\\}$ are alternately odd divisors of $u^\\alpha+1$ and  $u^\\alpha-1$.\n\\end{lemma}\n\n \\begin{lemma}{\\label{HIHLEW}}\\cite{HIHL}\nIf $G$ and $H$ are two non-abelian groups that satisfy $\\rho (G)\n~\\cap~ \\rho (H) = F$ with $|F| = n$, then $\\Delta (G\\times H) = K_n \\ast \\Delta (G) [\\rho (G) - F] \\ast \\Delta (H) [\\rho (H) - F]$, where $K_n$ is a complete graph with vertex set $F$.\n\\end{lemma}\n\\section{Proof of main results}\n\\noindent In this section, we wish to prove our main results.\\\\\n\n\\noindent\\textit{\\bf Proof of Theorem A.}\nOn the contrary, we assume that $\\Delta(G)$ is not perfect. Then by Lemma \\ref{hole}, there exists a subset $\\pi\\subseteq \\rho (G)$ such that for some integer $n\\geqslant 2$, $|\\pi|=2n+1$ and $\\Delta(G)^c[\\pi]$ or $\\Delta(G)[\\pi]$ is a cycle. Now one of the following cases occurs:\\\\\nCase 1. $\\Delta(G)^c[\\pi]$ is a cycle. Then there exist primes $p_0, p_1, \\dots , p_{2n} \\in \\rho(G)$ such that $\\Delta(G)^c[\\pi]$ is the cycle $C:p_0, p_1, \\dots , p_{2n},p_0$. Using Lemma \\ref{cycle},\n $N:=O^{\\pi^\\prime}(G)=S\\times A$, where $A$ is abelian, $S\\cong \\rm{SL}_2(p^m)$ or $S\\cong \\rm{PSL}_2(p^m)$ for a prime $p\\in \\pi$ and a positive integer $m$, and the primes in $\\pi - \\{p\\}$ are alternately odd divisors of $p^m+1$ and  $p^m-1$. Without loss of generality, we can assume that $p=p_0$. Since $\\Delta(G)^c[\\pi]$ is a cycle of length at least $5$, $p_1$ is not adjacent to both vertices $p_3$ and $p_4$ in $\\Delta(G)^c$. Also for some $\\epsilon \\in \\{\\pm 1\\}$, $p_3 \\in \\pi(p^m-\\epsilon)-\\{2\\}$ and $p_4 \\in \\pi(p^m+\\epsilon)-\\{2\\}$. Note that $p_1$ is an element of either $\\pi(p^m-\\epsilon)-\\{2\\}$ or $\\pi(p^m+\\epsilon)-\\{2\\}$.  Without loss of generality, we can assume that $p_1 \\in\\pi(p^m-\\epsilon)-\\{2\\}$. since $p_1$ and $p_4$ are adjacent vertices in $\\Delta(G)$, for some $\\chi \\in \\rm{Irr}(G)$, $p_1p_4|\\chi(1)$. Now let $\\varphi \\in \\rm{Irr}(N)$ be a constituent of $\\chi_N$. Then by Lemma \\ref{fraction}, $\\chi(1)/\\varphi(1)$ divides $[G:N]$. Therefore as $G/N$ is a $\\pi^\\prime$-group, $p_1p_4|\\varphi(1)$. It is a contradiction as using Lemma \\ref{chpsl}, $p_1$ and $p_4$ are non-adjacent vertices in $\\Delta(N)$.\\\\\nCase 2. $\\Delta(G)[\\pi]$ is a cycle. If $\\Delta(G)[\\pi]\\cong C_5$, then $\\Delta(G)^c[\\pi]\\cong C_5$ which is a contradiction with Case 1. Thus $|\\pi|\\geqslant 7$. Hence there exist distinct primes $p_1, p_2, p_3,q_1, q_2, q_3\\in \\pi$ such that the induced subgraphs of $\\Delta(G)^c$ on the sets  $\\pi_1:=\\{p_1, p_2, p_3\\}$ and $\\pi_2:=\\{q_1, q_2, q_3\\}$ are cycles of length $3$. Hence by Lemma \\ref{cycle}, for every $i=1,2$, $N_i:=O^{\\pi_i^\\prime}(G)=S_i\\times A_i$, where $A_i$ is abelian, $S_i\\cong \\rm{SL}_2(u_i^{\\alpha_i})$ or $S_i\\cong \\rm{PSL}_2(u_i^{\\alpha_i})$ for a prime $u_i\\in \\pi_i$ and a positive integer $\\alpha_i$, and the primes in $\\pi_i - \\{u_i\\}$ are alternately odd divisors of $u_i^{\\alpha_i}+1$ and  $u_i^{\\alpha_i}-1$. Let $N:=S_1S_2$. It is easy to see that $N/Z(N)\\cong \\rm{PSL}_2(u_1^{\\alpha_1})\\times\\rm{PSL}_2(u_2^{\\alpha_2})$. Therefore by Lemmas \\ref{chpsl} and \\ref{HIHLEW}, $\\Delta(N/Z(N))[\\pi_1 \\cup \\pi_2]\\subseteq \\Delta(G)$ is a complete bipartite graph with parts $\\pi_1$ and $\\pi_2$ which is a contradiction as $\\Delta(G)[\\pi]$ is a cycle of $\\Delta(G)$.\\qed\\\\\n\n\\noindent\\textit{\\bf Proof of Corollary B.}\nUsing Theorem A, $\\Delta(G)$ is a perfect graph. Thus by Lemma \\ref{comp}, $\\Delta(G)^c$ is too. Hence $\\chi(\\Delta(G)^c)=\\omega(\\Delta(G)^c)$. It is clear that $\\omega(\\Delta(G)^c)=\\alpha(\\Delta(G))$. By Lemma \\ref{pal}, $\\alpha(\\Delta(G))\\leqslant 3$. Hence $\\chi(\\Delta(G)^c)\\leqslant 3$ and the proof is completed.\\qed\n\n\\section*{Acknowledgements}\nThis research was supported in part\nby a grant  from School of Mathematics, Institute for Research in Fundamental Sciences (IPM).\n\n\n\n\n", "meta": {"timestamp": "2019-08-01T02:05:59", "yymm": "1907", "arxiv_id": "1907.13292", "language": "en", "url": "https://arxiv.org/abs/1907.13292", "abstract": "For a finite group $G$, let $\\Delta(G)$ denote the character graph built on the set of degrees of the irreducible complex characters of $G$. In graph theory, a perfect graph is a graph $\\Gamma$ in which the chromatic number of every induced subgraph $\\Delta$ of $\\Gamma$ equals the clique number of $\\Delta$. In this paper, we show that the character graph $\\Delta(G)$ of a finite group $G$ is always a perfect graph. We also prove that the chromatic number of the complement of $\\Delta(G)$ is at most three.", "subjects": "Group Theory (math.GR)", "title": "The character graph of a finite group is perfect", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9893474910448, "lm_q2_score": 0.855851143290548, "lm_q1q2_score": 0.8467341813223274}}
{"url": "https://arxiv.org/abs/0808.1065", "title": "Infinite log-concavity: developments and conjectures", "abstract": "Given a sequence (a_k) = a_0, a_1, a_2,... of real numbers, define a new sequence L(a_k) = (b_k) where b_k = a_k^2 - a_{k-1} a_{k+1}. So (a_k) is log-concave if and only if (b_k) is a nonnegative sequence. Call (a_k) \"infinitely log-concave\" if L^i(a_k) is nonnegative for all i >= 1. Boros and Moll conjectured that the rows of Pascal's triangle are infinitely log-concave. Using a computer and a stronger version of log-concavity, we prove their conjecture for the nth row for all n <= 1450. We also use our methods to give a simple proof of a recent result of Uminsky and Yeats about regions of infinite log-concavity. We investigate related questions about the columns of Pascal's triangle, q-analogues, symmetric functions, real-rooted polynomials, and Toeplitz matrices. In addition, we offer several conjectures.", "text": "\\section{Introduction}\n\nLet \n$$\n(a_k)=(a_k)_{k\\ge0}=a_0,a_1,a_2,\\ldots\n$$ \nbe a sequence of real numbers.  It will\nbe convenient to extend the sequence to negative indices by letting\n$a_k=0$ for $k<0$. Also, if $(a_k)=a_0,a_1,\\ldots,a_n$ is a finite\nsequence then we let $a_k=0$ for $k>n$.\n\nDefine the {\\it ${\\cal L}$-operator\\/} on sequences to be ${\\cal L}(a_k)=(b_k)$\nwhere $b_k=a_k^2-a_{k-1}a_{k+1}$.  \nCall a sequence {\\it $i$-fold log-concave\\/} if ${\\cal L}^i(a_k)$\nis a nonnegative sequence.  So log-concavity in the ordinary sense is\n$1$-fold log-concavity.\nLog-concave sequences arise in many areas of algebra, combinatorics,\nand geometry.  See the survey articles of Stanley~\\cite{sta:lus} and\nBrenti~\\cite{bre:lus} for more information.\n\nBoros and Moll~\\cite[page 157]{bm:ii} defined $(a_k)$ to be \n{\\it  infinitely log-concave\\/} if it is $i$-fold log-concave  for all\n$i\\ge1$.  They introduced this definition in conjunction with the\nstudy of a specialization of the Jacobi polynomials whose coefficient\nsequence they conjectured to be infinitely log-concave.   \nKauers and Paule~\\cite{kp:cpm} used a computer algebra package to prove\nthis conjecture for ordinary log-concavity.  \nSince the coefficients of these polynomials can be expressed in terms of\nbinomial coefficients, Boros and Moll also made the statement:\n\\begin{center}\n``Prove that the binomial coefficients are $\\infty$-logconcave.''\n\\end{center}\nWe will take this to be a conjecture that the rows of Pascal's triangle are infinitely log-concave, although we will later discuss the columns and other lines. \nWhen\ngiven a function of more than one variable, we will \nsubscript the ${\\cal L}$-operator by the parameter which is varying to form\nthe sequence.  \nSo ${\\cal L}_k\\binom{n}{k}$ would refer to the operator acting on the\nsequence $\\binom{n}{k}_{k\\ge0}$.  \nNote that we drop the sequence parentheses\nfor sequences of binomial coefficients to improve readability.\nWe now restate\nthe Boros-Moll conjecture formally.\n\\begin{conj}\n\\label{bm:con}\nThe sequence $\\binom{n}{k}_{k\\ge0}$ is  infinitely log-concave for all\n$n\\ge0$.\n\\end{conj}\n\nIn the next section, we use a strengthened version of log-concavity\nand computer calculations to verify Conjecture~\\ref{bm:con} for all\n$n\\le 1450$.  Uminsky and Yeats~\\cite{uy:uri} set up a correspondence\n between certain symmetric\nsequences and points in ${\\mathbb R}^m$.  They then described an infinite region \n${\\cal R}\\subset{\\mathbb R}^m$ bounded by hypersurfaces and such that\neach sequence corresponding to a point of ${\\cal R}$ \nis infinitely log-concave.  In Section~\\ref{ril}, we show how\nour methods can be used to give a simple derivation of one of their\nmain theorems.  We investigate infinite log-concavity of the columns\nand other lines of Pascal's \ntriangle in Section~\\ref{cpt}.  \nSection~\\ref{gps} is devoted to two $q$-analogues of the binomial\ncoefficients.  For the Gaussian polynomials, we show that certain\nanalogues of some infinite log-concavity conjectures \nare false while others appear to be true.  In contrast, our second\n$q$-analogue seems to retain all the log-concavity properties of the\nbinomial coefficients.   In Section~\\ref{sf}, after showing why the\nsequence $(h_k)_{k\\ge0}$ of complete homogeneous symmetric is an\nappropriate analogue of sequences of binomial coefficients, we explore\nits log-concavity properties.  We end with a section of related\nresults and questions about real-rooted polynomials and Toeplitz matrices.\n\nWhile one purpose of this article is to present our\n  results, we have written it with two more targets in mind.\n  The first is to convince our audience that infinite log-concavity is\n  a fundamental concept.  We hope that its definition as a natural\n  extension of traditional log-concavity helps to make this case. \nOur second aspiration is to attract others to work on the subject; to\n  that end, we have presented several open problems.  These\n  conjectures each represent fundamental questions in the area, so even\n  solutions of special cases may be interesting. \n\n\\section{Rows of Pascal's triangle}\n\\label{rpt}\n\nOne of the difficulties with proving the Boros-Moll conjecture is that\nlog-concavity is not preserved by the ${\\cal L}$-operator.  For example,\nthe sequence \n$4,5,4$ is log-concave but ${\\cal L}(4,5,4)=16,9,16$ is not.  So we\nwill seek a condition stronger than log-concavity which is preserved\nby ${\\cal L}$.  Given $r\\in{\\mathbb R}$, we say that a sequence $(a_k)$ is\n{\\it $r$-factor  log-concave\\/} if \n\\begin{equation}\n\\label{r}\na_k^2\\ge r a_{k-1}a_{k+1}\n\\end{equation}\nfor all $k$.  Clearly this implies log-concavity if $r\\ge1$.\n\nWe seek an $r>1$ such that $(a_k)$ being $r$-factor log-concave implies\nthat $(b_k)={\\cal L}(a_k)$ is as well.  Assume the original sequence is\nnonnegative.  Then expanding $r b_{k-1} b_{k+1}\\le b_k^2$ in terms of\nthe $a_k$ and rearranging the summands, we see that this is equivalent\nto proving\n$$\n(r-1) a_{k-1}^2 a_{k+1}^2 + 2 a_{k-1} a_k^2 a_{k+1}\n\\le a_k^4 + r  a_{k-2} a_k (a_{k+1}^2-a_k a_{k+2} ) + r a_{k-1}^2 a_k a_{k+2}.\n$$\n\nBy our assumptions, the two expressions with factors of $r$ on the right are\nnonnegative, so it suffices to prove the inequality obtained when\nthese are dropped.  Applying~\\ree{r} to the left-hand side gives\n$$\n(r-1) a_{k-1}^2 a_{k+1}^2 + 2 a_{k-1} a_k^2 a_{k+1}\n\\le  \\frac{r-1}{r^2} a_k^4 + \\frac{2}{r} a_k^4.\n$$\nSo we will be done if\n$$\n \\frac{r-1}{r^2}+ \\frac{2}{r} =1.\n$$\nFinding the root $r_0>1$ of the corresponding quadratic equation\nfinishes the proof of the first assertion of the following lemma, while the second assertion follows easily from the first.\n\n\\begin{lem}\n\\label{rfactor:le}\nLet \n$(a_k)$ be a nonnegative sequence and let $r_0=(3+\\sqrt{5})/2$.  \nThen $(a_k)$ being $r_0$-factor log-concave \nimplies that ${\\cal L}(a_k)$ is too. So in this case $(a_k)$ is infinitely\nlog-concave.\\hfill \\qed\n\\end{lem}\n\nNow to prove that any row of Pascal's triangle is infinitely\nlog-concave, one merely lets a computer find ${\\cal L}_k^i\\binom{n}{k}$ for\n$i$ up to some bound $I$.\nIf these sequences are all nonnegative and\n${\\cal L}_k^I\\binom{n}{k}$ is $r_0$-factor log-concave, then the\nprevious lemma shows that this row is infinitely log-concave.  Using\nthis technique, we have obtained the following theorem.\n\\begin{thm}\n\\label{row:th}\nThe sequence $\\binom{n}{k}_{k\\ge0}$ is infinitely log-concave for\nall $n\\le 1450$.\\hfill \\qed\n\\eth\n\nWe note that the necessary value of $I$ increases slowly with increasing $n$.\nAs an example, when $n=100$, our technique works with $I=5$, while for\n$n=1000$, we need $I=8$.\n\nOf course, the method developed in this section can be applied to any\nsequence such that ${\\cal L}^i(a_k)$ is $r_0$-factor log-concave for some\n$i$.  In particular, it is interesting to try it on the original\nsequence which motivated Boros and Moll~\\cite{bm:ii} to define infinite\nlog-concavity.  They were studying the polynomial\n\\begin{equation}\n\\label{Pm}\nP_m(x) = \\sum_{\\ell=0}^m d_\\ell(m) x^\\ell\n\\end{equation}\nwhere\n$$\nd_\\ell(m) = \\sum_{j=\\ell}^m 2^{j-2m} \n\\binom{2m-2j}{m-j}\\binom{m+j}{m}\\binom{j}{\\ell}.\n$$\nKauers~[private communication] has used our method to verify\ninfinite log-concavity of the sequence $(d_\\ell(m))_{\\ell\\ge0}$ for\n$m\\le 129$.  For such values of $m$, ${\\cal L}^5_\\ell$ applied to the sequence is \n$r_0$-factor log-concave.\n\n\\section{A region of infinite log-concavity}\n\\label{ril}\n\nUminsky and Yeats~\\cite{uy:uri}  took a different approach to the\nBoros-Moll Conjecture as described in the Introduction.  Since they\nwere motivated by the rows of Pascal's triangle, they only considered\nreal sequences $a_0,a_1,\\ldots,a_n$ which are symmetric (in that\n$a_k=a_{n-k}$ for all $k$) and satisfy $a_0=a_n=1$.  Each such sequence\ncorresponds to a point $(a_1,\\ldots,a_m)\\in{\\mathbb R}^m$ where $m=\\fl{n/2}$.\n\nTheir region, ${\\cal R}$, whose points all correspond to infinitely log-concave\nsequences, is bounded by $m$ parametrically defined hypersurfaces.  The\nparameters are $x$ and $d_1,d_2,\\ldots,d_m$  and it will be convenient\nto have the notation\n$$\ns_k=\\sum_{i=1}^k d_i.\n$$\nWe will also need $r_1=(1+\\sqrt{5})/2$.  Note that $r_1^2=r_0$.\nThe $k$th hypersurface, $1\\le k< m$, is defined as\n$$\n\\begin{array}{l}\n{\\cal H}_k=\n\\{(x^{s_1},\\ldots,x^{s_{k-1}},r_1x^{s_k},x^{s_{k+1}+d_k-d_{k+1}},\\ldots,x^{s_m+d_k-d_{k+1}}):\n\\\\[5pt]\n\\hs{100pt} x\\ge1, \\quad 1=d_1>\\cdots> d_k>d_{k+2}>\\cdots>d_m>0\\},\n\\end{array}\n$$\nwhile\n$$\n{\\cal H}_m=\n\\{(x^{s_1},\\ldots,x^{s_{m-1}},cx^{s_{m-1}}):\\\nx\\ge1, \\quad 1=d_1>\\cdots > d_{m-1}>0\\},\n$$\nwhere\n$$\nc=\\case{r_1}{if $n=2m$,}{2}{if $n=2m+1$.}\n$$\n\nLet us say that the \\emph{correct side} of ${\\cal H}_k$ for $1\\le k\\le m$ consists of those points in ${\\mathbb R}^m$ that can be obtained from a point on ${\\cal H}_k$ by increasing the $k$th coordinate.\nThen let ${\\cal R}$ be the region of all points in ${\\mathbb R}^m$ having\nincreasing coordinates and lying on the correct side of ${\\cal H}_k$ for all $k$.\nWe will show how our method of the previous section can be used to give\na simple proof  of one of  Uminsky and Yeats' main theorems.  But first we\nneed a modified version of Lemma~\\ref{rfactor:le} to take care of the\ncase when $n=2m+1$.\n\n\\begin{lem}\n\\label{rfactor:mod}\nLet $a_0,a_1,\\ldots, a_{2m+1}$ be a symmetric, nonnegative sequence such that\n\\begin{enumerate}\n\\item[(i)]  $a_k^2 \\ge r_0 a_{k-1} a_{k+1}$ for $ k< m$, and\n\\item[(ii)]  $a_m\\ge 2 a_{m-1}$.\n\\end{enumerate}\nThen ${\\cal L}(a_k)$ has the same properties, which implies that $(a_k)$ is\ninfinitely log-concave.\n\\end{lem}\n\\begin{proof}\nClearly ${\\cal L}(a_k)$ is still symmetric.  To show that the other\ntwo properties persist, note that in demonstrating\nLemma~\\ref{rfactor:le} we actually proved more.  In particular, we\nshowed that if equation~\\ree{r} holds at index $k$ of the \nsequence $(a_k)$ (with $r=r_0$), then it also holds at index $k$ of the\nsequence ${\\cal L}(a_k)$ provided that the original sequence is log-concave.\nNote that the assumptions of the current lemma imply log-concavity of\n$(a_k)$:  This is clear at indices $k\\neq m,m+1$ because of \ncondition~(i).  Also, using symmetry \nand multiplying condition (ii) by $a_m$ gives\n$a_m^2 \\ge 2 a_{m-1} a_{m} = 2 a_{m-1} a_{m+1}$ (and symmetrically for\n$k=m+1$).  \n\nSo now we know that condition~(i) is also true for ${\\cal L}(a_k)$.  As for\ncondition~(ii), using symmetry we see that we need to prove\n$$\na_m^2-a_{m-1} a_m\\ge 2\\left( a_{m-1}^2-a_{m-2} a_m \\right).\n$$\nRearranging terms and dropping one of them shows that it suffices to\ndemonstrate\n$$\n2 a_{m-1}^2 + a_{m-1} a_m \\le a_m^2.\n$$\nBut this is true because of~(ii), and we are done.\n\\end{proof}\n\n\n\\begin{thm}[\\cite{uy:uri}]\nAny sequence corresponding to a point of ${\\cal R}$ is infinitely log-concave.\n\\eth\n\\begin{proof}\nIt suffices to show that the sequence satisfies the hypotheses of \nLemma~\\ref{rfactor:le} when $n=2m$, or Lemma~\\ref{rfactor:mod} when $n=2m+1$.\n\nSuppose first that\n$k<m$.  Being on the correct side of ${\\cal H}_k$ \nis equivalent to there being values of the parameters such that\n$$\na_k^2\n\\hs{5pt}\\ge\\hs{5pt} \n(r_1 x^{s_k})^2\n\\hs{5pt}=\\hs{5pt}\nr_1^2 x^{(s_{k-1}+d_k)+(s_{k+1}-d_{k+1})}\n\\hs{5pt}=\\hs{5pt}\nr_0 a_{k-1} a_{k+1}.\n$$\nThus we have the necessary inequalities for this range of $k$.\n\nIf $k=m$ then we can use an argument as in the previous paragraph if $n=2m$.\nIf\n$n=2m+1$, then being on the correct side of ${\\cal H}_m$ \nis equivalent to\n$$\na_m\\ge 2 x^{s_{m-1}}=2 a_{m-1}.\n$$\nThis is precisely condition~(ii) of Lemma~\\ref{rfactor:mod}, which finishes\nthe proof.\n\\end{proof}\n\n\\section{Columns and other lines of Pascal's triangle} \n\\label{cpt}\n\nWhile we have treated Boros and Moll's statement about the infinite\nlog-concavity of the binomial coefficients to be a statement about the\nrows of Pascal's triangle, their wording also suggests an examination\nof the columns.\n\\begin{conj}\n\\label{col:con}\nThe sequence $\\binom{n}{k}_{n\\ge k}$ is  infinitely log-concave for\nall fixed $k\\ge0$. \n\\end{conj}\n\nWe will give two pieces of evidence for this conjecture.  One is a demonstration\nthat various columns corresponding to small values \nof $k$ are infinitely log-concave.\n Another is a proof that ${\\cal L}_n^i\\binom{n}{k}$ is\nnonnegative for certain values of $i$ and all $k$.\n\n\\begin{prop}\n\\label{small:k}\nThe sequence $\\binom{n}{k}_{n\\ge k}$ is  infinitely log-concave for\n$0\\le k \\le 2$.\n\\end{prop}\n\\begin{proof}  When $k=0$ we have, for all $i \\geq 1$, \n$$\n{\\cal L}_n^i\\binom{n}{0}=(1,0,0,0,\\ldots).\n$$\nFor $k=1$ we obtain\n$$\n{\\cal L}_n\\binom{n}{1} = (1,1,1,\\dots)\n$$\nso infinite log-concavity follows from the $k=0$ case.\nThe sequence when $k=2$ is a fixed point of\nthe ${\\cal L}$-operator, again implying infinite log-concavity.   \n\\end{proof}\n\nIn what follows, we use the notation $L(a_k)$ for the $k$th\nelement of the sequence ${\\cal L}(a_k)$, and similarly for $L_k$ and $L_n$.\n\n\\begin{prop}\n\\label{small:i}\nThe sequence ${\\cal L}_n^i\\binom{n}{k}$ is nonnegative for all $k$ and for\n$0\\le i\\le 4$.\n\\end{prop}\n\\begin{proof}\nBy the previous proposition, we only need to check $k\\ge3$.\nUsing the expression for a binomial coefficient in terms of\nfactorials, it is easy to derive the following expressions:\n$$\nL_n\\binom{n}{k} = \n\\frac{1}{n}\\binom{n}{k}\\binom{n}{k-1}\n$$\nand\n$$\nL_n^2\\binom{n}{k} = \n\\frac{2}{n^2 (n-1)} \n\\binom{n}{k}^2 \\binom{n}{k-1}\\binom{n}{k-2}.\n$$\n\nWith a little more work, one can show that $L_n^3\\binom{n}{k}$ can\nbe expressed as a product of nonnegative factors times the polynomial\n$$\n(4k-6)n^2-(4k^2-10k+6)n-k^2.\n$$ \nTo show that this is nonnegative, we write $n=k+m$ for $m\\ge0$ to get\n$$\n(4k-6)m^2+(4k^2-2k-6)m+(3k^2-6k).\n$$\nBut the coefficients of the powers of $m$ are all positive for $k\\ge3$,\nso we are done with the case $i=3$.\n\nWhen $i=4$, we follow the same procedure, only now the polynomial in\n$m$ has coefficients which are polynomials in $k$ up to degree $7$.\nFor example, the coefficient of $m^3$ is\n$$\n528 k^7- 8 k^6 - 11,248 k^5 + 25,360 k^4 - 5,888 k^3 - 24,296 k^2\n+16,080 k - 1,584. \n$$\nTo make sure this is nonnegative for integral $k\\ge3$, one rewrites\nthe polynomial as\n$$\n(528 k^2- 8 k - 11,248) k^5 + (25,360 k^2 - 5,888 k - 24,296) k^2\n+(16,080 k - 1,584), \n$$\nfinds the smallest $k$ such that each of the factors in parentheses is\nnonnegative from this value on, and then checks any remaining $k$ by\ndirect substitution. \n\\end{proof}\n\nKauers and Paule~\\cite{kp:cpm} proved that the rows of Pascal's\ntriangle are $i$-fold log-concave for $i\\le5$.\nKauers~[private communication] has used their techniques to confirm \nProposition~\\ref{small:i} and to also check the case $i=5$ for the\ncolumns.  For the latter case, Kauers used a computer to determine \n\\begin{equation} \n\\label{kau}\n\\frac{({\\cal L}_n^5 \\binom{n}{k})}{\\binom{n}{k}^{2^5}}\n\\end{equation}\nexplicitly, which \nis just a rational function in $n$ and $k$.  He then showed that  \n\\ree{kau} is nonnegative by means of cylindrical algebraic decomposition.\nWe refer the interested reader to~\\cite{kp:cpm} and the references therein for more\ninformation on such techniques.\n\nMore generally, we can look at an arbitrary line in Pascal's triangle,\ni.e.,  consider the sequence $\\binom{n+mu}{k+mv}_{m\\ge0}$.  \nThe unimodality and (1-fold) log-concavity of such sequences has been investigated in \n\\cite{bbs:ucs, sw:upp, tz:usb, tz:usb2}.\nWe do not require that $u$ and $v$ be coprime, so such sequences\nneed not contain all of the binomial coefficients in which a geometric\nline would intersect Pascal's triangle, e.g., a sequence such as \n$\\binom{n}{0}, \\binom{n}{2},\\binom{n}{4},\\ldots$ would be included.\nBy letting $u<0$, one can get a finite truncation of a column.  For\nexample, if $n=5$, $k=3$, $u=-1$, and $v=0$ then we get the sequence\n$$\n\\binom{5}{3}, \\binom{4}{3}, \\binom{3}{3}\n$$ \nwhich is not even $2$-fold log-concave.   So we will only consider $u\\ge0$.\nAlso\n$$\n\\binom{n+mu}{k+mv} = \\binom{n+mu}{n-k+m(u-v)}\n$$  \nso we can also assume $v\\ge0$.\n\nWe offer the following conjecture, which includes \nConjecture~\\ref{bm:con} as a special case.\n\\begin{conj}\n\\label{all:con}\nSuppose that $u$ and $v$ are distinct nonnegative integers.  Then\n$\\binom{n+mu}{mv}_{m\\ge0}$ is infinitely log-concave for all $n \\geq 0$ if\nand only if $u<v$ or $v=0$. \n\\end{conj}\n\nWe first give a quick proof of the ``only if'' direction.  Supposing\nthat $u>v \\ge1$, we consider the sequence \n$$\n\\binom{0}{0}, \\binom{u}{v}, \\binom{2u}{2v}, \\ldots\n$$\nobtained when $n=0$.\nWe claim that this sequence is not even log-concave and that\nlog-concavity fails at the second term.  Indeed, the fact that\n$\\binom{u}{v}^2 < \\binom{2u}{2v}$ follows immediately from the\nidentity\n\\[\n\\binom{u}{0}\\binom{u}{2v} + \\binom{u}{1}\\binom{u}{2v-1} + \\cdots + \n\\binom{u}{v}\\binom{u}{v} + \\cdots + \\binom{u}{2v}\\binom{u}{0} = \\binom{2u}{2v},\n\\]\nwhich is a special case of Vandermonde's Convolution.\n\nThe proof just given shows that subsequences of the columns of\nPascal's triangle are the only {\\it infinite} sequences of the form\n$\\binom{n+mu}{mv}_{m\\ge0}$ that can possibly be infinitely log-concave.   \nWe also note that the previous conjecture says nothing about what\nhappens on the diagonal $u=v$.  Of course, the case $u=v=1$ is\nConjecture~\\ref{col:con}.  For other diagonal values, the evidence is\nconflicting. \nOne can show by computer that $\\binom{n+mu}{mu}_{m\\ge0}$ is not $4$-fold log-concave\nfor $n=2$ and any\n$2\\le u\\le500$.  However, this is the only {\\it known} value of $n$\nfor which $\\binom{n+mu}{mu}_{m\\ge0}$ is not an infinitely log-concave\nsequence for some $u\\ge1$. \n\nWe conclude this section by offering considerable computational\nevidence in favor of the ``if'' direction of Conjecture~\\ref{all:con}.\nTheorem~\\ref{row:th} provides such evidence when $u=0$ and $v=1$.  Since all\nother\nsequences with $u<v$ have a finite number of nonzero entries, we can\nuse the $r_0$-factor log-concavity technique for these sequences as well.\nFor all $n\\le 500$, $2\\le v\\le20$ and $0\\le u<v$, we have checked that\n$\\binom{n+mu}{mv}_{m\\ge0}$ is infinitely log-concave.\n\n\\section{$q$-analogues}\n\\label{gps}\n\nThis section will be devoted to discussing two $q$-analogues of\nbinomial coefficients.   For the Gaussian polynomials, we will see\nthat the corresponding generalization of Conjecture~\\ref{bm:con} is\nfalse, and we show one exact reason why it fails.  In contrast, the\ncorresponding generalization of Conjecture~\\ref{col:con} appears to be\ntrue.   This shows how delicate these conjectures are and may in part\nexplain why they seem to be difficult to prove.  After introducing our\nsecond $q$-analogue, we\nconjecture that the corresponding generalizations of\nConjectures~\\ref{bm:con}, \\ref{col:con} and \\ref{all:con} are all\ntrue.  This second $q$-analogue arises in the study of quantum groups; see,\nfor example, the books of Jantzen~\\cite{jan:lqg} and Majid~\\cite{maj:qgp}.\n\n\nLet $q$ be a variable and consider a polynomial $f(q)\\in{\\mathbb R}[q]$.\nCall $f(q)$ \\linebreak $q$-{\\it nonnegative\\/}  if all the coefficients of $f(q)$ are\nnonnegative.  Apply\nthe ${\\cal L}$-operator to sequences of polynomials $(f_k(q))$ in the\nobvious way.  Call such a sequence {\\it $q$-log-concave\\/} if\n${\\cal L}(f_k(q))$ is a sequence of $q$-nonnegative polynomials, with\n$i$-fold $q$-log-concavity and infinite $q$-log-concavity defined\nsimilarly.  \n\nWe will be particularly interested in the Gaussian polynomials.  The\nstandard {\\it $q$-analogue of the nonnegative integer $n$\\/} is\n$$\n[n]=[n]_q=\\frac{1-q^n}{1-q} = 1+q+q^2+\\cdots+q^{n-1}.\n$$\nThen, for $0\\le k\\le n$, the {\\it Gaussian polynomials\\/} or \n{\\it $q$-binomial coefficients\\/} are defined as\n$$\n\\gaus{n}{k}=\\gaus{n}{k}_q = \\frac{[n]_q!}{[k]_q! [n-k]_q!}\n$$\nwhere $[n]_q!=[1]_q [2]_q\\cdots [n]_q$.   For more information,\nincluding proofs of the assertions made in the next paragraph, see the\nbook of Andrews~\\cite{and:tp}.\n\nClearly substituting $q=1$ gives $\\genfrac{[}{]}{0pt}{}{n}{k}_1=\\binom{n}{k}$.\nAlso, it is well known that the Gaussian polynomials\nare indeed $q$-nonnegative polynomials.  In fact, they have\nvarious combinatorial interpretations, one of which we will need.\nAn (integer) {\\it partition of $n$\\/} is a weakly decreasing\npositive integer sequence $\\lambda=(\\lambda_1,\\lambda_2,\\ldots,\\lambda_\\ell)$ such\nthat $|\\lambda|\\stackrel{\\rm def}{=}\\sum_i \\lambda_i=n$.  The $\\lambda_i$ are\ncalled {\\it parts\\/}.  For notational convenience, if a part $k$ is repeated $r$ times in a\npartition $\\lambda$ then we will denote this by writing $k^r$ in the\nsequence for $\\lambda$.\nWe say that $\\lambda$ {\\it fits inside an $s\\times t$ box} if\n$\\lambda_1\\le t$ and  $\\ell\\le s$.  Denote the set of all such partitions\nby $P(s,t)$. \nIt is well known, and easy to prove by induction on $n$, that\n\\begin{equation}\n\\label{gau}\n\\gaus{n}{k}=\\sum_{\\lambda\\in P(n-k,k)} q^{|\\lambda|}.\n\\end{equation}\n\nWe are almost ready to prove that the sequence\n$\\left(\\gaus{n}{k}\\right)_{k\\ge0}$ is not infinitely $q$-log-concave.\nIn fact, we will show it is not even $2$-fold $q$-log-concave.\nFirst we need a lemma.  In it, we use $\\mathop{\\rm mint}\\nolimits f(q)$ to denote the\nnonzero term of least degree in $f(q)$.\n\\begin{lem}\n\\label{mint}\nLet $L_k\\left(\\gaus{n}{k}\\right)=B_k(q)$.  Then for $k\\le n/2$,\n$$\n\\mathop{\\rm mint}\\nolimits B_k(q)=\n\\case{q^k}{if $k<n/2$,}{2q^k}{if $k=n/2$.}\n$$\n\\end{lem}\n\\begin{proof}\nSince $B_k(q)=\\gaus{n}{k}^2-\\gaus{n}{k-1}\\gaus{n}{k+1}$ it suffices to\nprove, in view of~\\ree{gau}, the following two statements.  If $i\\le k$ and \n$$\n(\\lambda,\\mu)\\in P(n-k+1,k-1)\\times P(n-k-1,k+1)\n$$ \nwith $|\\lambda|+|\\mu|=i$, then $(\\lambda,\\mu)\\in P(n-k,k)^2$.  Furthermore,\nthe number of elements in \n$P(n-k,k)^2- P(n-k+1,k-1)\\times P(n-k-1,k+1)$ is 0 or 1 or 2 depending\non whether $i<k$ or $i=k<n/2$ or $i=k=n/2$, respectively.\n\nThe first statement is an easy consequence of \n$|\\lambda|+|\\mu|=i\\le k\\le n-k$.  A similar argument works for the $i<k$\ncase of the second statement.  If $i=k$ then the pair $((k),\\emptyset)$ is in the\ndifference and if $i=k=n/2$ then the pair $(\\emptyset,(1^k))$ is as well.\n\\end{proof}\n\n\\begin{prop}\n\\label{gau:pr}\nLet $L_k^2\\left(\\gaus{n}{k}\\right)=C_k(q)$.  Then for\n$n\\ge2$ and $k=\\fl{n/2}$ we have\n$$\n\\mathop{\\rm mint}\\nolimits C_k(q)=-q^{n-2}.\n$$\nConsequently, $\\left(\\gaus{n}{k}\\right)_{k\\ge0}$ is not $2$-fold $q$-log-concave.\n\\end{prop}\n\\begin{proof}\nThe proofs for $n$ even and odd are similar, so we will only do the\nformer.  So suppose $n=2k$ and consider \n$$\nC_k(q)=B_k(q)^2-B_{k-1}(q)B_{k+1}(q)=B_k(q)^2-B_{k-1}(q)^2.\n$$\nBy the previous lemma\n$\\mathop{\\rm mint}\\nolimits B_k(q)^2=4q^{2k}$ and $\\mathop{\\rm mint}\\nolimits B_{k-1}(q)^2= q^{2k-2}$.  Thus\n$\\mathop{\\rm mint}\\nolimits C_k(q) = -q^{2k-2}=-q^{n-2}$ as desired.\n\\end{proof}\n\nAfter what we have just proved, it may seem surprising that the\nfollowing conjecture, which is a $q$-analogue of\nConjecture~\\ref{col:con}, does seem to hold. \n\\begin{conj}\n\\label{qcol:con}\nThe sequence $\\left(\\gaus{n}{k}\\right)_{n\\ge k}$ is  infinitely\n$q$-log-concave for\nall fixed $k\\ge0$. \n\\end{conj}\n\nAs evidence, we will prove a $q$-analogue of Proposition~\\ref{small:k}\nand comment on Proposition~\\ref{small:i} in this setting.\n\\begin{prop}\n\\label{qsmall:k}\nThe sequence $\\left(\\gaus{n}{k}\\right)_{n\\ge k}$ is  infinitely\n$q$-log-concave for\n$0\\le k\\le 2$.\n\\end{prop}\n\\begin{proof}\nWhen $k=0$ one has the same sequence as when $q=1$.  \n\nWhen $k=1$ we claim that\n$$\n{\\cal L}\\left(\\gaus{n}{1}\\right) = (1,q,q^2,q^3,\\ldots).\n$$\nIndeed, \n\n\\begin{eqnarray*}\n[n]^2-[n-1][n+1]\n&=& \\frac{(1-q^n)^2-(1-q^{n-1})(1-q^{n+1})}{(1-q)^2}\\\\\n&=& \\frac{q^{n-1}-2q^n+q^{n+1}}{(1-q)^2}\\\\\n&=& q^{n-1}\n\\end{eqnarray*}\n(and recall that the sequence starts at $n=1$).  It follows that\n$$\n{\\cal L}^i\\left(\\gaus{n}{1}\\right) = (1,0,0,0,\\ldots)\n$$\nfor $i\\ge2$.\n\nFor $k=2$, the manipulations are much like those in the previous paragraph.\nUsing induction on $i$, we obtain\n$$\nL^i\\left(\\gaus{n}{2}\\right) = q^{(2^i-1)(n-2)}\\gaus{n}{2}\n$$\nfor $i\\ge0$.  This completes the proof of the last case of the\nproposition.  \n\\end{proof}\n\nIf we now consider arbitrary $k$ it is not hard to show, using\nalgebraic manipulations like those in the proof just given,  \nthat\n\\begin{equation}\n\\label{lq}\nL_n \\left(\\gaus{n}{k}\\right)=\n\\frac{q^{n-k}}{[n]}\\gaus{n}{k}\\gaus{n}{k-1}.\n\\end{equation}\nThese are, up to a power of $q$, the $q$-Narayana numbers.  They were\nintroduced by F\\\"urlinger and Hofbauer~\\cite{fh:qcn} and are contained\nin a specialization of a result of MacMahon~\\cite[page 1429]{mac:cp}\nwhich was stated without proof.  They were further studied by\nBr\\\"and\\'en~\\cite{bra:qnn}.  As shown in the references just cited,\nthese polynomials are the generating functions for a number of\ndifferent families of combinatorial objects.  Thus they are $q$-nonnegative.\n\nMore computations show that\n\\begin{equation}\n\\label{llq}\nL_n^2\\left(\\gaus{n}{k}\\right)\n=\\frac{q^{3n-3k}[2]}{[n]^2[n-1]}\\gaus{n}{k}^2\\gaus{n}{k-1}\\gaus{n}{k-1}.\n\\end{equation}\nIt is not clear that these polynomials are $q$-nonnegative, although they\nmust be if Conjecture~\\ref{qcol:con} is true.  Furthermore, when $q=1$,\nthe triangle made as $n$ and $k$ vary is not in Sloane's\nEncyclopedia~\\cite{slo:oei} (although it has now been submitted).  We\nexpect that these integers and polynomials have interesting, yet to be\ndiscovered, properties.\n\n\nWe conclude our discussion of the Gaussian polynomials by considering the\nsequence \n\\begin{equation}\n\\label{gs}\n\\left(\\gaus{n+mu}{mv}\\right)_{m\\ge0}\n\\end{equation}\nfor nonnegative integers $u$ and $v$, as we did in Section~\\ref{cpt} for the\nbinomial coefficients.\nWhen $u>v$ the sequence has an infinite number of nonzero entries.  We\ncan use $\\eqref{gau}$ to show that the highest degree term in\n$\\gaus{n+u}{v}^2 - \\gaus{n+2u}{2v}$ has coefficient $-1$, so the\nsequence \\eqref{gs} is not even $q$-log-concave.   When $u<v$, it\nseems to be the case that the sequence is not $2$-fold $q$-log-concave,\nas shown for the rows in Proposition~\\ref{gau:pr}.  When $u=v$, the\nevidence is conflicting, reflecting the behavior of the binomial\ncoefficients.  Since setting $q=1$ in $\\gaus{n+mu}{mu}$ yields \n$\\binom{n+mu}{mu}$, we know that\n$\\left(\\gaus{2+mu}{mu}\\right)_{m\\ge0}$ is not always \n$4$-fold $q$-log-concave.  It also transpires that the case $n=3$ is\nnot always $5$-fold $q$-log-concave.  We have not encountered other\nvalues of $n$ that fail to yield a $q$-log-concave sequence when\n$u=v$. \n\nWhile the variety of behavior of the Gaussian polynomials is\ninteresting, it would be desirable to have a $q$-analogue that\nbetter reflects the behavior of the binomial coefficients.  A\n$q$-analogue that arises in the study of quantum groups serves\nthis purpose.  Let us \nreplace the previous $q$-analogue of the nonnegative integer $n$ \nwith the expression \n$$\n\\langle n\\rangle = \\frac{q^n-q^{-n}}{q-q^{-1}} =\nq^{1-n} + q^{3-n} + q^{5-n} + \\cdots + q^{n-1}.\n$$\nFrom this, we obtain a $q$-analogue of the binomial coefficients by\nproceeding as for the Gaussian polynomials: for $0\\le k\\le n$, we\ndefine \n$$\n\\qua{n}{k} = \\frac{\\langle n\\rangle!}{\\langle k\\rangle! \\langle n-k\\rangle!}\n$$\nwhere $\\langle n\\rangle!=\\langle1 \\rangle \\langle2 \\rangle\\cdots \\langle\nn\\rangle$. \n\nLetting $q\\to1$ in $\\qua{n}{k}$ gives $\\binom{n}{k}$, and a\nstraightforward calculation shows that  \n\\begin{equation}\n\\label{quagau}\n\\qua{n}{k} = \\frac{1}{q^{nk-k^2}}\\gaus{n}{k}_{q^2}.\n\\end{equation}\nSo $\\qua{n}{k}$ is, in general, a Laurent polynomial in $q$\nwith nonnegative coefficients.  Our definitions of $q$-nonnegativity and $q$-log-concavity for polynomials in $q$ extend to Laurent polynomials in the obvious way.\n\nWe offer the following generalizations of Conjectures~\\ref{bm:con},\n\\ref{col:con} and \\ref{all:con}. \n\n\\begin{conj}\n\\label{qua:con}\n\\\n\\begin{itemize} \n\\item[(a)] The row sequence $\\left(\\qua{n}{k}\\right)_{k\\ge0}$ is\n  infinitely $q$-log-concave for all $n\\ge0$. \n\\item[(b)] The column sequence $\\left(\\qua{n}{k}\\right)_{n\\ge\n    k}$ is infinitely $q$-log-concave for all fixed $k\\ge0$. \n\\item[(c)] For all integers $0\\le u<v$, the sequence\n  $\\left(\\qua{n+mu}{mv}\\right)_{m\\ge0}$ is infinitely\n  $q$-log-concave for all $n\\ge0$. \n\\end{itemize}\n\\end{conj}\n\nSeveral remarks are in order.  Suppose that for $f(g), g(q) \\in\n{\\mathbb R}[q, q^{-1}]$, we say $f(q) \\le g(q)$  if $g(q)-f(q)$ is $q$-nonnegative.\nThen the proofs of Lemmas~\\ref{rfactor:le} and\n\\ref{rfactor:mod} work equally well if the $a_i$'s are Laurent\npolynomials and we replace the term ``log-concave'' by\n``$q$-log-concave.''  Using these lemmas, we have verified\nConjecture~\\ref{qua:con}(a) for all $n\\le 53$.   Even though (a) is a\nspecial case of (c), we state it separately since (a) is the\n$q$-generalization of the Boros-Moll conjecture, the primary\nmotivation for this paper.  \n\nAs evidence for Conjecture~\\ref{qua:con}(b), it is not hard to prove\nthe appropriate analogue of Propositions~\\ref{small:k} and\n\\ref{qsmall:k}, i.e.\\ that the sequence $\\qua{n}{k}_{n\\ge k}$ is\ninfinitely $q$-log-concave for all $0\\le k\\le2$.  To obtain the\nexpressions for $L_n\\left(\\qua{n}{k}\\right)$ and\n$L_n^2\\left(\\qua{n}{k}\\right)$, take\nequations~\\eqref{lq} and \\eqref{llq}, replace all square\nbrackets by angle brackets and replace each the terms $q^{n-k}$ and\n$q^{3n-3k}$ by the number 1. \n\nConjecture~\\ref{qua:con}(c) has been verified for all $n\\le 24$\nwith $v\\le10$.   When $u > v$, we can use $\\eqref{quagau}$ to show\nthat the lowest degree term in  $\\qua{n+u}{v}^2 - \\qua{n+2u}{2v}$ has\ncoefficient $-1$, so the sequence is not even $q$-log-concave. \nWhen $u=v$, the quantum groups analogue has exactly the same behavior\nas we observed above for the Gaussian polynomials. \n\n\\section{Symmetric functions}\n\\label{sf}\n\nWe now turn our attention to symmetric functions.  \nWe will demonstrate that the complete homogeneous symmetric functions\n$(h_k)_{k\\ge0}$ are a natural analogue of the rows and columns of\nPascal's triangle.  We show that the sequence\n$(h_k)_{k\\ge0}$ is \n$i$-fold log-concave in the appropriate sense for $i\\le3$, \nbut not $4$-fold log-concave.\nLike the results of Section~\\ref{gps}, this\nresult underlines the difficulties and subtleties of\nConjectures~\\ref{bm:con} and~\\ref{col:con}.  In particular, it shows\nthat any proof of Conjecture~\\ref{bm:con} or Conjecture~\\ref{col:con}\nwould need to use techniques that do not carry over to the sequence\n$(h_k)_{k\\ge0}$.   \nFor a more detailed\nexposition of the background material below, we refer the reader to\nthe texts of Fulton~\\cite{ful:yt}, Macdonald~\\cite{mac:sfh},\nSagan~\\cite{sag:sym} or Stanley~\\cite{sta:ec2}.  \n\n\nLet ${\\bf x}=\\{x_1,x_2,\\ldots\\}$ be a countably infinite set of variables.\nFor each $n\\ge0$, the elements of the symmetric group $\\mathfrak{S}_n$ act on formal\npower series $f({\\bf x})\\in{\\mathbb R}[[{\\bf x}]]$ by permutation of variables (where\n$x_i$ is left fixed if $i>n$).  The algebra of symmetric functions,\n$\\Lambda({\\bf x})$, is the set of all series left fixed by all symmetric\ngroups and of bounded (total) degree.  \n\nThe vector space of symmetric functions homogeneous of degree $k$ has\ndimension equal to the number of partitions\n$\\lambda=(\\lambda_1,\\ldots,\\lambda_\\ell)$ of $k$.  We will be \ninterested in three bases for this vector space.  The \n{\\it monomial symmetric function\\/} corresponding to $\\lambda$,\n$m_\\lambda=m_\\lambda({\\bf x})$, is obtained by \nsymmetrizing the monomial $x_1^{\\lambda_1}\\cdots x_\\ell^{\\lambda_\\ell}$.  The\n$k$th {\\it complete homogeneous symmetric function\\/}, $h_k$, is the sum of\nall monomials of degree $k$.  For partitions, we then define \n$$\nh_\\lambda=h_{\\lambda_1}\\cdots h_{\\lambda_\\ell}.\n$$  \nFinally, the \n{\\it Schur function\\/} corresponding to $\\lambda$ is\n$$\ns_\\lambda= \\det(h_{\\lambda_i-i+j})_{1\\le i,j \\le \\ell}.\n$$\nWe remark that this determinant is a minor of the Toeplitz matrix for\nthe sequence $(h_k)$.  We will have more to say about Toeplitz matrices\nin the next section.\n\nOur interest will be in the sequence just mentioned\n$(h_k)_{k\\ge0}$.  Let $h_k(1^n)$ \ndenote the integer obtained by substituting $x_1=\\cdots=x_n=1$ and\n$x_i=0$ for $i>n$ into $h_k=h_k({\\bf x})$.  Then $h_k(1^n)=\\binom{n+k-1}{k}$\n(the number of ways of choosing $k$ things from $n$ things with\nrepetition) and so the above sequence becomes a column of Pascal's\ntriangle.  By the same token $h_k(1^{n-k})=\\binom{n-1}{k}$ and so the\nsequence becomes a row.\n\nWe will now collect the results from the theory of symmetric functions\nwhich we will need.  \nPartially order partitions by {\\it dominance\\/} \nwhere $\\lambda\\le\\mu$ if and only if for every $i\\ge1$ we have\n$\\lambda_1+\\cdots+\\lambda_i\\le\\mu_1+\\cdots+\\mu_i$.  Also,  if  $\\{b_\\lambda\\}$ is\nany basis of $\\Lambda({\\bf x})$ and $f\\in\\Lambda({\\bf x})$ then we let\n$[b_\\lambda]f$ denote the coefficient of the basis element $b_\\lambda$ in the\nexpansion of $f$ in this basis.   First we have a simple consequence\nof Young's Rule.\n\\begin{thm}\n\\label{yr}\nFor any partitions $\\lambda,\\mu$ we have $[m_\\mu] s_\\lambda$ is a nonnegative\ninteger.  In particular,\n\n\\vs{5pt}\n\n\\eqed{\n[m_\\mu]s_\\lambda=\n\\case{$1$}{if $\\mu=\\lambda$,}\n{$0$}{if $\\mu\\not\\le \\lambda.$}\n}\n\\eth\n\nLet $\\lambda+\\mu$ denote the componentwise sum\n$(\\lambda_1+\\mu_1,\\lambda_2+\\mu_2,\\ldots)$.  \nThe next result follows from the\nLittlewood-Richardson Rule and induction.\n\\begin{thm}\n\\label{lrr}\nFor any partitions $\\lambda^1,\\ldots,\\lambda^r$ and $\\mu$ we have\n$[s_\\mu] s_{\\lambda^1}\\cdots s_{\\lambda_r}$ is a nonnegative integer.  In particular,\n\n\\vs{5pt}\n\n\\eqed{\n[s_\\mu]s_{\\lambda^1}\\cdots s_{\\lambda^r}=\n\\case{$1$}{if $\\mu=\\lambda^1+\\cdots+\\lambda^r$,}\n{$0$}{if $\\mu\\not\\le \\lambda^1+\\cdots+\\lambda^r$.}\n}\n\\eth\nBecause of this result we call $\\lambda^1+\\cdots+\\lambda^r$ the \n{\\it dominant partition\\/} for $s_{\\lambda^1}\\cdots s_{\\lambda^r}$.\n\nFinally, we need a result of Kirillov~\\cite{kir:csg} about the product\nof Schur functions, which was proved bijectively by Kleber~\\cite{kle:prs} and\nFulmek and Kleber~\\cite{fk:bps}.  This result can be obtained by applying the Desnanot-Jacobi Identity---also known as Dodgson's condensation formula---to the Jacobi-Trudi matrix for $s_{k^{r+1}}$.\nNote that, to improve readability, we drop the sequence parentheses when a sequence appears as a subscript.\n\n\\begin{thm}[\\cite{fk:bps,kir:csg,kle:prs}]\n\\label{kir:thm}\nFor positive integers $k,r$ we have\n\n\\vs{5pt}\n\n\\eqed{\n\\left(s_{k^r}\\right)^2 - s_{(k-1)^r} s_{(k+1)^r} = s_{k^{r-1}} s_{k^{r+1}}.\n}\n\\eth\n\nTo state our results, we need a few more definitions.  If ${b_\\lambda}$ is\na basis for $\\Lambda({\\bf x})$ and $f\\in\\Lambda({\\bf x})$ then we \nsay $f$ is {\\it $b_\\lambda$-nonnegative\\/} if $[b_\\lambda]f\\ge0$ for all\npartitions $\\lambda$.  Note that $m_\\lambda$-nonnegativity is the natural\ngeneralization to many variables of the $q$-nonnegativity definition for\n${\\mathbb R}[q]$.  Also note that $s_\\lambda$-nonnegativity implies\n$m_\\lambda$-nonnegativity by Theorem~\\ref{yr}. \n\\begin{thm}\nThe sequence ${\\cal L}^i(h_k)$ is $s_\\lambda$-nonnegative  for $0\\le  i\\le 3$.  \nBut the sequence ${\\cal L}^4(h_k)$ is not $m_\\lambda$-nonnegative.\n\\eth\n\\begin{proof}\nFrom the definition of the Schur function we have\n$$\nL^0(h_k)=h_k=s_k \\quad\\mbox{and}\\quad L^1(h_k)=(h_k)^2-h_{k-1}h_{k+1}=s_{k^2}.\n$$\nNow Theorem~\\ref{kir:thm} immediately gives\n$$\nL^2(h_k) = \\left(s_{k^2}\\right)^2 - s_{(k-1)^2}s_{(k+1)^2} = s_k s_{k^3}\n$$\nwhich is $s_\\lambda$-nonnegative by the first part of Theorem~\\ref{lrr}.\nUsing Theorem~\\ref{kir:thm} twice gives\n\\begin{eqnarray*}\nL^3(h_k)\n&=& \\left(s_k\\right)^2 \\left(s_{k^3}\\right)^2 - s_{k-1} s_{(k-1)^3} s_{k+1} s_{(k+1)^3}\\\\\n&=& \\left(s_k\\right)^2 \\left(s_{k^3}\\right)^2 - \\left(s_k\\right)^2 s_{(k-1)^3} s_{(k+1)^3}\\\\\n& & + \\left(s_k\\right)^2 s_{(k-1)^3} s_{(k+1)^3} - s_{k-1} s_{(k-1)^3} s_{k+1} s_{(k+1)^3}\\\\\n&=& \\left(s_k\\right)^2 s_{k^2} s_{k^4} +  s_{(k-1)^3} s_{k^2} s_{(k+1)^3}\n\\end{eqnarray*}\nwhich is again $s_\\lambda$-nonnegative.  This finishes the cases \n$0\\le i\\le3$.\n\nWe now assume $k\\ge2$.  Computing $L^4(h_k)$ from the expression for\n$L^3(h_k)$ gives the sum of the terms in the left column below.  The right column gives the dominant partition for each term, as determined by Theorem~\\ref{lrr}.\n\\[\n\\begin{array}{ll}\n+(s_k)^4 (s_{k^2})^2  (s_{k^4})^2 \n& (8k, 4k, 2k, 2k) \\\\\n+2 (s_k)^2 (s_{k^2}) ^2  s_{k^4} s_{(k-1)^3} s_{(k+1)^3} \n& (7k, 5k, 3k, k) \\\\\n+ (s_{(k-1)^3})^2 (s_{k^2})^2  (s_{(k+1)^3})^2\n& (6k, 6k, 4k) \\\\\n-  (s_{k-1})^2 s_{(k-1)^2} s_{(k-1)^4} (s_{k+1})^2 s_{(k+1)^2} s_{(k+1)^4} \n& (8k, 4k, 2k, 2k) \\\\\n- (s_{k-1})^2 s_{(k-1)^2} s_{(k-1)^4} s_{k^3} s_{(k+1)^2} s_{(k+2)^3} \n& (7k-1, 5k+1, 3k+1, k-1) \\\\\n-s_{(k-2)^3} s_{(k-1)^2} s_{k^3} (s_{k+1})^2 s_{(k+1)^2} s_{(k+1)^4} \n& (7k+1, 5k-1, 3k-1, k+1)\\\\\n-s_{(k-2)^3} s_{(k-1)^2} (s_{k^3})^2 s_{(k+1)^2} s_{(k+2)^3} \n& (6k, 6k, 4k)\n\\end{array}\n\\]\n\nNow consider $\\lambda=(7k+1,5k-1,3k-1,k+1)$, the dominant partition for the penultimate term above.  Observe that if $\\mu$ is the dominant partition for any other\nterm, then $\\lambda\\not\\le\\mu$.  So, by the\nsecond part of Theorem~\\ref{lrr}, $s_\\lambda$ appears in the Schur-basis\nexpansion for $L^4(h_k)$ with coefficient $-1$.  It then follows from the \nsecond part of Theorem~\\ref{yr}, that\nthe coefficient of $m_\\lambda$ is $-1$ as well.\n\\end{proof}\n\n\n\\section{Real roots and Toeplitz matrices}\n\\label{rrp}\n\nWe now consider two other (almost equivalent) settings where, in contrast\nto the results of the previous section, Conjecture~\\ref{bm:con} does\nseem to generalize.  In fact, this may be the right level of\ngenerality to find a proof.\n\nLet $(a_k)=a_0,a_1,\\ldots,a_n$ be a finite sequence of nonnegative real numbers.\nIt was shown by Isaac Newton that if all the roots of the polynomial\n$p[a_k]\\stackrel{\\rm def}{=}a_0 + a_1 x + \\cdots a_n x^n$ are real, then the\nsequence $(a_k)$\nis log-concave.  For example, since the polynomial\n$(1+x)^n$ has only real roots, the $n$th row of Pascal's triangle is\nlog-concave.  It is natural to ask if the real-rootedness property is \npreserved by the ${\\cal L}$-operator.\nThe literature includes a\nnumber of results about operations on polynomials which preserve\nreal-rootedness; for example,\nsee~\\cite{bra:ltp,bre:ulp,bre:lus,pit:pbc,wag:tph,wy:prz}.\n\n\\begin{conj}\n\\label{pla:con}\nLet $(a_k)$ be a finite sequence of nonnegative real numbers.  If $p[a_k]$\nhas only real roots then the same is true of $p[{\\cal L}(a_k)]$.\n\\end{conj}\n\nThis conjecture is due independently to \nRichard Stanley~[private communication].  It is also one of a number\nof related conjectures made by Steve Fisk~\\cite{fis:qdp}.\nIf true, Conjecture~\\ref{pla:con}\nwould immediately imply the original Boros-Moll Conjecture.\nAs evidence for the conjecture, we have verified it by computer for a\nlarge number of randomly chosen real-rooted polynomials.  We have also checked\nthat\n$p[{\\cal L}_k^i\\binom{n}{k}]$ has only real roots for all $i\\le 10$ and\n$n\\le 40$.  It is interesting to note that  Boros and Moll's\npolynomial $P_m(x)$ in equation~\\ree{Pm} does not have real roots even\nfor $m=2$.  So if the corresponding sequence is infinitely log-concave\nthen it must be so for some other reason.\n\nAlong with the rows of Pascal's triangle, it appears that applying ${\\cal L}$ to the other finite lines we were considering in Section~\\ref{cpt} also yields sequences with real-rooted\ngenerating functions.\nSo we make the following conjecture\nwhich implies the ``if'' direction of Conjecture~\\ref{all:con}.\n\\begin{conj}\nFor $0\\le u<v$, the polynomial $p[{\\cal L}_m^i(\\binom{n+mu}{mv})]$  has\nonly real roots for all $i\\ge0$.  \n\\end{conj}\nWe have verified this assertion for all $n\\le24$ with $i\\le10$ and\n$v\\le10$.  \nIn fact, it follows from a theorem of Yu~\\cite{yu:ctc} that the conjecture holds for $i=0$ and all $0\\le u<v$.  So it will suffice to prove Conjecture~\\ref{pla:con} to obtain this result for all $i$.\n\nWe can obtain a matrix-theoretic perspective on problems of\nreal-rooted\\-ness via \nthe following renowned result of Aissen, Schoenberg and\nWhitney~\\cite{asw:gft}.  A matrix $A$ is said to be totally\nnonnegative if every minor of $A$ is nonnegative.  \nWe can associate with any sequence $(a_k)$ a corresponding\n(infinite) {\\it Toeplitz matrix\\/} $A = (a_{j-i})_{i,j\\ge0}$.\nIn comparing the\nnext theorem to Newton's result, note that for a real-rooted\npolynomial $p[a_k]$ the roots being nonpositive is\nequivalent to the sequence $(a_k)$ being nonnegative.\n\n\\begin{thm}[\\cite{asw:gft}]\n\\label{asw:th}\nLet $(a_k)$ be a finite sequence of real numbers.  Then every root of\n$p[a_k]$ is a nonpositive real number if and only if the \nToeplitz matrix $(a_{j-i})_{i,j\\ge0}$ is totally nonnegative. \\hfill \\qed\n\\eth\n\nTo make a connection with the ${\\cal L}$-operator, note that\n$$\na_k^2-a_{k-1}a_{k+1}= \\left| \\begin{array}{cc} a_k & a_{k+1} \\\\ a_{k-1} & a_k\n\\end{array} \\right|,\n$$\nwhich is a minor of the Toeplitz matrix \n$A =(a_{j-i})_{i,j\\ge0}$.  Call such a minor {\\it adjacent}\nsince its entries are adjacent in $A$.    \nNow, for an arbitrary infinite matrix \n$A = (a_{i,j})_{i,j\\geq 0}$, let us define the infinite matrix ${\\cal L}(A)$ by \n$$\n{\\cal L}(A) = \\left( \\left| \n\\begin{array}{cc} a_{i,j}   & a_{i,j+1} \\\\\n                  a_{i+1,j} & a_{i+1,j+1} \n\\end{array} \n\\right| \\right)_{i,j \\ge0}. \n$$\nNote that if $A$ is the Toeplitz matrix of $(a_k)$ then ${\\cal L}(A)$ is\nthe Toeplitz matrix of ${\\cal L}(a_k)$.\nUsing Theorem~\\ref{asw:th}, Conjecture~\\ref{pla:con} can now be strengthened as\nfollows.\n\n\\begin{conj}\n\\label{tnn:con}\nFor a sequence $(a_k)$ of real numbers,\nif $A = (a_{j-i})_{i,j\\ge0}$ is totally nonnegative then ${\\cal L}(A)$ is also\ntotally nonnegative.\n\\end{conj}\n\nNote that if $(a_k)$ is finite, then Conjecture~\\ref{tnn:con} is equivalent to Conjecture~\\ref{pla:con}.\nAs regards evidence for Conjecture~\\ref{tnn:con}, consider an arbitrary $n$-by-$n$\nmatrix $A = (a_{i,j})_{i,j=1}^n$.  For finite matrices, ${\\cal L}(A)$ is\ndefined in the obvious way to be the $(n-1)$-by-$(n-1)$ matrix\nconsisting of the 2-by-2 adjacent minors of $A$.  \nIn~\\cite[Theorem 6.5]{fhgj:ctp}, Fallat, Herman, Gekhtman, and Johnson\nshow that for $n\\le4$, ${\\cal L}(A)$\nis totally nonnegative whenever $A$ is.  However,\nfor $n=5$, an example from their paper can be modified to show that\nif \n$$\nA = \n\\left(\n\\begin{array}{ccccc}\n1 & t & 0 & 0 & 0 \\\\\nt & t^2+1 & 2t & t^2 & 0\\\\\nt^2 & t^3+2t & 1+4t^2 & 2t^3+t & 0\\\\\n0 & t^2 & 2t^3+2t & t^4+2t^2+1 & t\\\\\n0 & 0 & t^2 & t^3+t & t^2\n\\end{array}\n\\right)\n$$\nthen $A$ is totally nonnegative for $t\\ge0$,  but ${\\cal L}(A)$ is not\ntotally nonnegative for sufficiently large $t$ ($t\\ge\\sqrt{2}$ will\nsuffice).  We conclude that the Toeplitz structure would be important\nto any affirmative answer to Conjecture~\\ref{tnn:con}. \n\nWe finish our discussion of the matrix-theoretic perspective with a\npositive result similar in flavor to Conjecture~\\ref{tnn:con}. \n\n\\begin{prop}\nIf $A$ is a finite square matrix that is positive semidefinite, then ${\\cal L}(A)$\nis also positive semidefinite.\n\\end{prop}\n\n\\begin{proof}\nThe key idea is to construct the {\\it second compound matrix}\n$\\mathcal{C}_2(A)$ of $A$, which is the array of all 2-by-2 minors of\n$A$, arranged lexicographically according to the row and column\nindices of the minors~\\cite{hj:ma}.  \n\nWe claim that if $A$ is\npositive semidefinite, then so is $\\mathcal{C}_2(A)$.  Indeed, since\nthe compound operation preserves multiplication and inverses, the\neigenvalues of $\\mathcal{C}_2(A)$ are equal to the eigenvalues of\n$\\mathcal{C}_2(J)$, where $J$ is the Jordan form of $A$.  If $J$ is\nupper-triangular and has diagonal entries $\\lambda_1, \\lambda_2, \\ldots,\n\\lambda_n$, then we see that $\\mathcal{C}_2(J)$ is upper-triangular with\ndiagonal entries $\\lambda_i \\lambda_j$ for all $i<j$.  Since the $\\lambda_i$'s are\nall nonnegative, so too are the eigenvalues of $\\mathcal{C}_2(J)$,\nimplying that $\\mathcal{C}_2(A)$ is positive semidefinite.   \n\nFinally, since ${\\cal L}(A)$ is a principal submatrix of\n$\\mathcal{C}_2(A)$, ${\\cal L}(A)$ is itself positive semidefinite.   \n\\end{proof}\n\n\\medskip\n\n{\\it Acknowledgements.}  We thank Bodo Lass for suggesting that we\napproach Conjecture~\\ref{bm:con} from the point-of-view of real roots\nof polynomials.  Section~\\ref{rrp} also benefited from interesting\ndiscussions with Charles R. Johnson. \n\n", "meta": {"timestamp": "2009-03-24T21:38:29", "yymm": "0808", "arxiv_id": "0808.1065", "language": "en", "url": "https://arxiv.org/abs/0808.1065", "abstract": "Given a sequence (a_k) = a_0, a_1, a_2,... of real numbers, define a new sequence L(a_k) = (b_k) where b_k = a_k^2 - a_{k-1} a_{k+1}. So (a_k) is log-concave if and only if (b_k) is a nonnegative sequence. Call (a_k) \"infinitely log-concave\" if L^i(a_k) is nonnegative for all i >= 1. Boros and Moll conjectured that the rows of Pascal's triangle are infinitely log-concave. Using a computer and a stronger version of log-concavity, we prove their conjecture for the nth row for all n <= 1450. We also use our methods to give a simple proof of a recent result of Uminsky and Yeats about regions of infinite log-concavity. We investigate related questions about the columns of Pascal's triangle, q-analogues, symmetric functions, real-rooted polynomials, and Toeplitz matrices. In addition, we offer several conjectures.", "subjects": "Combinatorics (math.CO)", "title": "Infinite log-concavity: developments and conjectures", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9891815526228442, "lm_q2_score": 0.8558511488056151, "lm_q1q2_score": 0.8465921681895833}}
{"url": "https://arxiv.org/abs/1905.10483", "title": "On the product dimension of clique factors", "abstract": "The product dimension of a graph $G$ is the minimum possible number of proper vertex colorings of $G$ so that for every pair $u,v$ of non-adjacent vertices there is at least one coloring in which $u$ and $v$ have the same color. What is the product dimension $Q(s,r)$ of the vertex disjoint union of $r$ cliques, each of size $s$? Lovász, Nešetřil and Pultr proved in 1980 that for $s=2$ it is $(1+o(1)) \\log_2 r$ and raised the problem of estimating this function for larger values of $s$. We show that for every fixed $s$, the answer is still $(1+o(1)) \\log_2 r$ where the $o(1)$ term tends to $0$ as $r$ tends to infinity, but the problem of determining the asymptotic behavior of $Q(s,r)$ when $s$ and $r$ grow together remains open. The proof combines linear algebraic tools with the method of Gargano, Körner, and Vaccaro on Sperner capacities of directed graphs.", "text": "\\section{Introduction}\n\nThe product dimension of a graph $G=(V,E)$ is the minimum possible\ncardinality $d$ of a collection of proper vertex colorings of $G$ such\nthat every pair of nonadjacent vertices have the same color in at\nleast one of the colorings. Equivalently, this is the minimum $d$ so\nthat one can assign to every vertex $v$ a vector \nin $Z^d$, so that two vertices are adjacent if and only\nif the corresponding vectors differ in all coordinates. This\ndimension is also the minimum number of complete graphs so that $G$\nis an induced subgraph of their tensor product, \nwhere the tensor product of graphs $H_1, \\ldots ,H_d$ is the \ngraph whose vertex set is the cartesian product of the vertex sets of the graphs $H_i$, and two vertices $(u_1,u_2, \\ldots ,u_d)$ and $(v_1,v_2, \\ldots ,v_d)$ are adjacent iff $u_i$ is adjacent (in $H_i$)  to $v_i$ for all $1 \\leq i \\leq d$.\nYet another equivalent definition is the minimum number of subgraphs of the\ncomplement $\\overline{G}$ of $G$ so that each subgraph is a vertex disjoint \nunion of cliques, and every edge of $\\overline{G}$ belongs to at least one of the subgraphs.\n\nFor positive integers $s,r \\geq 2$ let $K_s(r)$ denote the graph\nconsisting of $r$ pairwise vertex disjoint copies of the complete\ngraph $K_s$, and let $Q(s,r)$ denote the product dimension\nof this graph. Lov\\'asz, Ne\\v{s}et\\v{r}il and Pultr \\cite{LNP}\n(see also \\cite{Al1})\nproved that $Q(2,r)=\\lceil \\log_2 (2r) \\rceil$.\nThe proof of the upper bound  is simple. If $q=\\lceil \\log_2 (2r) \n\\rceil$ then \n$2^q \\geq 2r$. Hence one can assign\ndistinct binary vectors of length $q$ to the $2r$ vertices of $K_2(r)$ \nso that the vectors assigned to each pair of adjacent vertices are\nantipodal, that is, differ in all coordinates. It is easy to check\nthat two vertices are adjacent if and only if the corresponding vectors\ndiffer in all coordinates, showing that $Q(2,r) \\leq q.$\n\nThe lower bound is proved in \\cite{LNP} by a linear algebra\nargument, and the proof given in \\cite{Al1} applies exterior\nalgebra. There is yet another (similar) short proof that proceeds\nby assigning to each vertex of $K_s(r)$ a multilinear polynomial \nin $x_1,x_2, \\ldots ,x_q$ that depends on the coloring used, and by showing\nthat these polynomials are linearly independent. \nAs mentioned in the abstract, \nLov\\'asz, Ne\\v{s}et\\v{r}il and Pultr \\cite{LNP} raised the problem\nof estimating $Q(s,r)$ for larger values of $s$. \nMore recently, Kleinberg and Weinberg considered the same problem, motivated by  the investigation of prophet inequalities for intersection of matroids \\cite{KW}. In this paper, we determine\nthe asymptotic behavior of $Q(s,r)$ for any fixed $s \\geq 2$ and large $r$.\n\\begin{theo}\n\\label{t11}\nFor every fixed $s$, $Q(s,r) = (1+o(1)) \\log_2 r$, where the \n$o(1)$-term tends to $0$ as $r$ tends to infinity.\n\\end{theo}\nThe main tool in the proof is the method of Gargano, K\\\"orner and\nVaccaro in their work on Sperner capacities \\cite{GKV}. \nFor completeness, and\nsince we are interested\nin the behavior of the $o(1)$-term in the theorem above, we\ndescribe a variant of the method as needed here, in a combinatorial \nway that avoids any application of information theoretic techniques.\nThe proof is based on what we call here $Z_s$-covering families of\nvectors. \n\nLet $Z_s$ denote the ring of integers  modulo $s$. For a subset\n$A \\subset Z_s$ and a vector $v=(v_1,v_2, \\ldots ,v_q) \\in Z_s^q$, \nwe say that $v$ is $A$-covering if for every $a \\in A$ there is an\n$1 \\leq i \\leq q$ so that $v_i=a$. The vector $v$ is covering if it\nis $Z_s$-covering. A family ${\\cal F} \\subset Z_s^q$ is $A$-covering if\nfor every ordered pair of  distinct vectors $u,v \\in {\\cal F}$, the difference\n$u-v$ is $A$-covering. ${\\cal F}$  is covering if it is $Z_s$-covering.\nTherefore, a family ${\\cal F}$ of vectors  in $Z_s^q$ is \ncovering if every element of $Z_s$ \nappears in at least one coordinate of \nthe  difference between any two distinct vectors in the\nfamily. Let $R(s,q)$ denote the maximum possible cardinality of\na covering family of vectors in $Z_s^q$. The following simple\nstatement describes the connection between $Q(s,r)$ and $R(s,q)$.\n\\begin{prop}\n\\label{p12}\nIf $R(s,q) \\geq r$ then $Q(s,r) \\leq q$.\n\\end{prop}\nNote that by definition $R(s,q)=1$ for all $q<s$. \nOur main result about $R(s,q)$ is the following.\n\\begin{theo}\n\\label{t13}\n\\begin{enumerate}\n\\item\nFor every $q \\geq s \\geq 2$, $R(s,q) \\leq 2^{q-1}$. Equality\nholds for $s=2$.\n\\item\nFor every fixed $s$, $R(s,q) \\geq (2-o(1))^q$, where the\n$o(1)$-term tends to $0$ as $q \\to \\infty$.\n\\end{enumerate}\n\\end{theo}\n\nThe rest of this paper is organized as follows. Section 2 contains\nthe proof of Proposition \\ref{p12} and that of a simple combinatorial lemma.\nIn Section 3 we present the proof of Theorem \\ref{t13} and note that\nin view of Proposition \\ref{p12}\nit implies Theorem \\ref{t11}. The proof supplies\nbetter estimates for prime values of $s$, and we thus first present\nthe proof for this special case (which suffices to deduce the assertion of\nTheorem \\ref{t11} for every $s$) and then \ndescribe briefly the \nproof for general $s$. The final Section 4 contains some \nconcluding remarks and open problems, including some (modest) estimates for\n$R(s,q)$ when $q$ is not much larger than $s$.\n\nTo simplify the presentation we omit, throughout the paper, all \nfloor and ceiling signs whenever these are not crucial.\n\n\\section{Preliminaries}\n\nWe first prove Proposition \\ref{p12}.  If $R(s,q) \\ge r$, then there \nexists a matrix of elements in $Z_s$ with $r$ rows and $q$ columns so \nthat the difference of any two rows is covering.  We will use the $q$ \ncolumns of this matrix to find $q$ graphs, each being \na disjoint union of cliques, that cover the complement of $K_s(r)$.  This complement is a complete multipartite graph with $r$ parts of size $s$.  Label the vertices of this graph with elements of $Z_s$ so that in each part all labels are used exactly once. We will associate each row of the matrix to a part and each column to a vertex disjoint union of cliques.  \nFor a column $(a_1, \\cdots, a_r)^T$, consider the following graph.  For each $0 \\le k<s$ take the $k+a_i$th vertex (taken modulo $s$) from the $i$th part of size $s$ and take the union of the $s$ cliques obtained as $0 \\le k<s$ ranges.  This is clearly a union of $s$ vertex disjoint cliques.  \nNow, suppose we have some two vertices of the graph we are trying to cover in different parts, say the $\\ell+m$th vertex of part $i$ \nand the $\\ell$th vertex of part $j$ for some $1 \\le i<j \\le q$ \nand some $\\ell, m \\in Z_s$.  Then the difference of the $i$th and $j$th rows contains an $m$ in some column $(a_1, \\cdots, a_r)^T$, so that $a_i-a_j \\equiv m$ in $Z_s$, and then the disjoint union of cliques corresponding to this column will cover our desired edge.\n\nNext, we need the following simple lemma.\n\\begin{lemma}\n\\label{l21}\nLet $H$ be a bipartite graph with classes of vertices $A_1$, $A_2$ where\n$|A_1|=n_1$, $|A_2|=n_2$, each vertex of $A_1$ has degree $d_1$, and each vertex of $A_2$ has degree $d_2$. Furthermore \nsuppose that $d_2 \\ge \\log(2n_2)$.  Then there is a union of \nvertex-disjoint\nstars with centers in $A_1$, each star having at least\n$\\frac{d_1}{4 \\log (2n_2)}$ leaves, such that all vertices of\n$A_2$ are leaves.\n\\end{lemma}\n\n\\begin{proof}\n\tDefine a random subset $S$ of $A_1$ by choosing each vertex of $A_1$ to be in $S$ with probability $p=\\frac{\\log(2n_2)}{d_2}$ uniformly and independently.  We claim that with positive probability, each vertex of $A_2$ has between $1$ and $4\\log(2n_2)$ neighbors in $S$.  \nThe proof is a simple union bound; a fixed vertex $v \\in A_2$ has probability $$(1-p)^{d_2}< e^{-pd_2}=\\frac{1}{2n_2}$$ of having no neighbors in $S$, and probability at most $$\\dbinom{d_2}{4\\log(2n_2)}p^{4\\log(2n_2)} \\le \\left( \\frac{ped_2}{4\\log(2n_2)} \\right)^{4\\log(2n_2)}=\\left(\\frac{e}{4}\\right)^{4\\log(2n_2)}<\\frac{1}{2n_2}$$ of \nhaving more than $4\\log(2n_2)$ neighbors, proving the claim.  \nFix an $S$ with this property.\n\t\nWe now finish the proof of the lemma by an application of Hall's theorem.  \nFor all $S' \\subseteq S$, let $N(S')$ denote the set of all neighbors of\n$S'$ and let $e(S',A_2)$ denote the number of edges from $S'$ to\n$A_2$. Then\n$|N(S')| \\ge \\frac{e(S',A_2)}{4\\log(2n_2)}\n=\\frac{d_1}{4\\log(2n_2)}|S'|$.  \nHence every subset of $S$ \nexpands by a factor of at least \n$\\frac{d_1}{4\\log(2n_2)}$.  \nThus by Hall's theorem, there is a \nunion of disjoint stars whose centers are exactly \nthe vertices of $S$, each having at least $\\frac{d_1}{4\\log(2n_2)}$ leaves.  \nEvery remaining vertex of $A_2$ is adjacent to some vertex in $S$, \nso we can simply add it to an existing star.  \\end{proof}\n\n\\section{Covering families}\n\n\\subsection{The upper bound}\n\nThe following proposition implies the assertion of Theorem \\ref{t13},\npart 1.\n\\begin{prop}\n\\label{p31}\nFix $s \\geq 2$, and let ${\\cal F} \\subset Z_s^q$ be a $\\{0,1\\}$-covering\nfamily of vectors. Then $|{\\cal F}| \\leq 2^{q-1}$. For $s=2$ equality holds.\n\\end{prop}\n\\begin{proof}\nPut $m =|{\\cal F}|$. Let $p$ be a prime divisor of $s$ and consider the \nvectors in ${\\cal F}$ as vectors in $Z_p^q$ by reducing their coordinates modulo\n$p$. Note that these vectors form a $\\{0,1\\}$-covering family over $Z_p$. Let\n$v_i=(v_{i1},v_{i2}, \\ldots ,v_{iq})$, $(1 \\leq i \n\\leq m)$ be the vectors in ${\\cal F}$ (considered as elements of \n$Z_p^q$). \n\nFor each $1\\leq i \\leq m$ define two polynomials $P_i,Q_i$ in\nthe variables $x_1,x_2, \\ldots ,x_q$ over $Z_p$ as follows.\n$$\nP_i(x_1, \\ldots ,x_q)=\\prod_{j=1}^q (x_j -v_{ij}),~~\nQ_i(x_1, \\ldots ,x_q)=\\prod_{j=1}^q (x_j -v_{ij}-1).\n$$\nIt is not difficult to check that for every $i$, $Q_i(v_i) \\neq 0$\nand $P_i(v_i)=0$. In addition, for every \n$1 \\leq i \\neq i' \\leq m$, $P_{i'}(v_i) =0$ (as there is a coordinate\n$j$ for which $v_{ij}-v_{i'j}=0$) and $Q_{i'}(v_i)=0$ (as there\nis a $j$ so that $v_{ij}-v_{i'j}=1$).\n\nSimilar reasoning gives that for the vectors $v_i+J$, where $J$ is the all\n$1$-vector of length $q$, $P_i(v_i+J) \\neq 0$, $Q_i(v_i+J)=0$, and\nfor every $i' \\neq i$, \n$P_ {i'}(v_i+J) =Q_{i'} (v_i +J)=0$. Therefore, for each member of\nthe collection\nof $2m$ polynomials $\\{P_i,Q_i: 1 \\leq i \\leq m\\}$ there is an \nassignment of values of the variables in which this member is nonzero\nand all others vanish. This easily implies that the set of $2m$ polynomials\n$P_i,Q_i$ is linearly independent in $Z_p$, \nand as each of its members lies in the space of multilinear\npolynomials with the $m$ variables $x_j$, the number, $2m$,\nof these polynomials is at most the dimension of this space which is\n$2^q$. It follows that $|{\\cal F}| =m \\leq 2^{q-1}$, as needed. For $s=2$\nthe family of all binary vectors in which the first coordinate is $1$\nis $\\{0,1\\}$-covering, showing that $R(2,q)  \\geq 2^{q-1}$ and \ncompleting the proof. $\\Box$\n\n\\end{proof}\n\n\\subsection{Prime $s$} For prime $s \\ge 3$, we will prove that $R(s,q) \n\\ge (2-o_s(1))^q$ where the $o(1)$-term tends to $0$ as $q \\to \\infty$.  \nThe crux of the proof is a Markov chain argument from \\cite{GKV}, which we will iterate $O(\\log s)$ times.\n\nA \\emph{balanced word} over $Z_s^q$ is a word containing the letters $1$ through $s-1$ an equal number of times.  A \\emph{special balanced word} is such a balanced word so that the first $\\frac{q}{(s-1)/2}$ letters are $1$ and $2$ in some order, the next $\\frac{q}{(s-1)/2}$ letters are $3$ and $4$ in some order, and so on.  \nConstruct a bipartite graph between the set $A_2$ of balanced words \n$w$ over $Z_s^q$ and the set $A_1$ of permutations $\\pi$ on $q$ elements defined as follows: $w$ and $\\pi$ are adjacent if and only if $\\pi(w)$ is a special balanced word.  \nBy symmetry all vertices in $A_1$ have the same degree $d_1$, and all vertices in $A_2$ have the same degree $d_2$.  We have $$n_2=|A_2|=\\dbinom{q}{q/(s-1), \\cdots, q/(s-1)} \\le (s-1)^q$$ and $d_1$ is the total number of special balanced words, \nso $$d_1=\\dbinom{2q/(s-1)}{q/(s-1)}^{(s-1)/2}>\\frac{2^q}{q^{s/2}}.$$  Furthermore, $d_2=\\left((\\frac{q}{s-1})!\\right)^{s-1}d_1 \\ge \\log(2n_2)$ so by Lemma \\ref{l21} there exists a way to map balanced words to some set $T$ of permutations $\\pi$ of $q$ elements, so that each balanced word is \nassociated to exactly one permutation, and each permutation in $T$ is associated to at least $\\frac{d_1}{4\\log(2n_2)}>\\frac{2^q/q^{s/2}}{4\\log(2(s-1)^q)}>\\frac{2^q}{q^s}$ balanced words.  \nThus, we can partition the balanced words into sets $S_1, S_2, \\cdots$ so that for each $S_i$ we have $|S_i|>\\frac{2^q}{q^s}$ and for all $i$ there exists $\\pi_i$ so that $\\pi_i(s_i)$ is a special balanced word for all $s_i \\in S_i$.\n\nGiven all of the special balanced words of length $q$, any two of them have a difference vector which covers $\\{\\pm 1\\}$.  The idea of the proof will be to amplify this set $\\{-1,1\\}$, first to $\\{-\\alpha,-1,1,\\alpha\\}$ for a primitive root $\\alpha$ modulo $s$, and after $r$ steps to $\\{\\pm \\alpha^b\\}$ for $0 \\le b<2^r$.  At each stage, the number of vectors will be $(2-o(1))^L$ where $L$ is the length of the vectors.  Thus after $O(\\log s)$ steps we will have a set of vectors that is $Z_s^{*}$ covering.  We can then add an extra coordinate of $0$ to all of the vectors to make them $Z_s$-covering.\n\nWe describe the first step of this iteration in detail.  Fix a primitive root $\\alpha$ modulo $s$, which will be constant throughout the steps.  Also fix $n=\\frac{100}{\\epsilon}(\\log(s))^2$, which will again be constant throughout the steps.  \nInitially we set $q=q_0=(1+o(1))100s^2$, ensuring it is divisible by\n$s-1$.  \nWe will construct words of length $q_1=qn$ by stringing together balanced words of length $q$ in a specific way.  If $x_i \\in S_j$, then force $x_{i+1} \\in \\alpha S_j$; here we mean that if we take $x_{i+1}$ and multiply its letters by $\\alpha^{-1}$ pointwise, the result will be in $S_j$.  \nConsider all vectors of length $qn$ constructed according to this rule, by concatenating $n$ balanced words of length $q$ in this way.  \nThere are more than $n_2(\\frac{2^q}{q^s})^{n-1}$ such words, \nbecause at each stage other than the first we must pick \n$x_{i+1}$ so that $x_i \\in \\alpha S_j$ for some $j$, and thus there \nare more than $\\frac{2^q}{q^s}$ choices for $x_{i+1}$. \nWe will make $x_1, x_n$ uniform over all such words; \nthis costs us a factor of $n_2^2$ and thus we now \nhave more than $\\frac{1}{n_2}(\\frac{2^q}{q^s})^{n-1}$ such words.  \nUsing that $n_2<s^q$, we can find more than $\\frac{2^{q(n-1)}}{q^{s(n-1)}s^q}=(2-o(1))^{qn}$ words of length $qn$ of the form $x_1x_2 \\cdots x_n$ so that $x_1, x_n$ are fixed.  We now note that for any two different words of this form $x_1 \\cdots x_n$ and $x'_1 \\cdots x'_n$, \nwith $x_1=x'_1, x_n=x'_n$, there is some minimal $i \\ge 2$ so that $x_i \\neq x'_i$.  But then there is some $k$ so that $x_{i-1}=x'_{i-1} \\in S_k$, and that means $x_i, x'_i \\in \\alpha S_k$.  \nBecause $x_i \\neq x'_i$, we have $\\alpha^{-1}x_i \\neq \\alpha^{-1}x'_i$, and both $\\alpha^{-1}x_i$  and $\\alpha^{-1}x'_i$ are in $S_k$.  It follows that $\\alpha^{-1}x_i-\\alpha^{-1}x'_i$ covers $\\{\\pm 1\\}$ and so $x_i-x'_i$ covers $\\{\\pm \\alpha\\}$.  \nSimilarly, there is some maximal $i<n$ so that $x_i \\neq x'_i$.  Then $x_{i+1}=x'_{i+1} \\in \\alpha S_j$ for some $j$, so $x_i \\in S_j$, $x'_i \\in S_j$.  As $x_i \\neq x'_i$, it follows that $x_i-x'_i$ must cover $\\{\\pm 1\\}$ as coordinates.  \nWhen $s=5$, setting $\\alpha=2$ and applying this construction already \ngives a covering family for $\\mathbb{Z}_s^{*}=\\mathbb{Z}_5^*$ \nwith $(2-o(1))^N$ vectors of length $N$.  In the general case we iterate this argument to find $(2-o(1))^N$ vectors of length $N$, so that after the $r$th iteration the vectors we get cover $\\{\\pm \\alpha^b\\}$ for all $0 \\le b<2^r$.  We describe how to do this inductively.  \n\nAfter $r$ iterations, we find for some $q=q_r$ (that depends on $s$) \na family of $M^q=(2-o_q(1))^q$ (balanced) vectors over $Z_s^q$ which covers $\\pm \\alpha^b$ for all $0 \\le b<2^r$.  We call these vectors $y_1, \\cdots, y_{M^q}$ and let $Y=\\{y_1, \\cdots, y_{M^q}\\}$.  We repeat the above argument.  \nThe $y_i$ play the role of the special balanced words.  \nAgain we construct a bipartite graph.  \nOn one side there is a set $A_2=Y$ and on the \nother side there is $A_1$, permutations of $q$ elements.  \nWe have $\\pi$ is adjacent to a (balanced) vector $y$ if and only \nif $\\pi(y) \\in Y$.  Again $n_2=|A_2| \\le (s-1)^q$ but now $d_1=M^q$.  \nIt can easily be verified that $d_2 \\ge \\log(2n_2)$.  Thus by Lemma \\ref{l21} the balanced words of length $q$ will be split into sets $S_1, S_2, \\cdots$ so that all $S_i$ satisfy $|S_i| \\ge \\frac{d_1}{4\\log(2n_2)} \\ge \\frac{M^q}{4+4q\\log(s-1)}=(2-o_q(1))^q$, and furthermore for each $S_i$, there exists a permutation $\\pi_i$ so that for all $s_i \\in S_i$, $\\pi_i(S_i) \\in Y$.  Now we will again construct a Markov chain.  \nConsider all words of length $qn=q_{r+1}$ consisting of $n$ balanced words of length $q$ of the form $x_1x_2 \\cdots x_n$, so that if $x_i \\in S_j$, then $x_{i+1} \\in \\alpha^{2^r}S_j$.  If we have two such words $x_1x_2 \\cdots x_n$ and $x'_1x'_2 \\cdots x'_n$ so that $x_1=x'_1$, $x_n=x'_n$, let $i>1$ be minimal so that $x_i \\neq x'_i$.  Then if $x_{i-1}=x'_{i-1} \\in S_j$, $x_i, x'_i \\in \\alpha^{2^r}S_j$.  \nThis means that $x_i-x'_i$ covers $\\{\\pm \\alpha^b\\}$ modulo $s$ for $2^r \\le b<2^{r+1}$.  Now, if we let $x_i$, $i<n$, be maximal so that $x_i \\neq x'_i$, then $x_{i+1}=x'_{i+1} \\in \\alpha^{2^r}S_j$ for some $j$ and so $x_i, x'_i$ are not equal but are both in $S_j$.  \nThis means $x_i-x'_i$ covers $\\{\\pm \\alpha^b\\}$ modulo $s$ \nfor $0 \\le b<2^r$.  So indeed the family covers $\\{\\pm \\alpha^b\\}$ \nmodulo $s$ \nfor $0 \\le b<2^{r+1}$, as long as $x_1$ and $x_n$ are fixed over the family.  We can always find \n$$\n\\frac{\\min_{i}|S_i|^{n-1}}{\\dbinom{q}{q/(s-1), \\cdots, q/(s-1)}} \n\\ge \\frac{(2-o(1))^{qn}}{(s-1)^q}=(2-o(1))^{qn}\n$$ \nsuch words if $n$ is large enough, where the $o(1)$ terms tend to $0$ \nas $q \\to \\infty$ and $n$ is sufficiently large.\nWe will soon see that our choice of $n=\\frac{100}{\\epsilon}(\\log(s))^2$ \nis sufficient.  Iterating the argument $ \\log_2(s) $ \ntimes allows us to find $(2-o(1))^q$ vectors of some length $q$ which cover $Z_s^{*}$.  \nAdding a single coordinate where all vectors are $0$ gives us vectors of \nlength $q+1$ that cover $Z_s$, without changing the asymptotic analysis.\n\nWith care, we can extract quantitative bounds.  We assume $\\epsilon>0$ is a fixed constant and show that we only require $q=\\left(\\frac{O(1)}{\\epsilon}(\\log(s))^2\\right)^{\\log(s)}$ to have a $(2-\\epsilon)^q$ size covering system over $Z_s^q$ for large $s$.  \nSay that after $r$ iterations, we have $M^q=M_r^q$ vectors of length \n$q=q_r$ for some $M \\le 2$ (which is nearly $2$).  \nThen $\\min_{i}S_i \\ge \\frac{M^q}{5q\\log(s)}$  and thus we find at least $$\\frac{\\min_{i}|S_i|^{n-1}}{\\dbinom{q}{q/(s-1), \\cdots, q/(s-1)}} \\ge \\frac{M^{qn}}{M^q(5q\\log(s))^ns^q} \\ge \\frac{M^{qn}}{(2s)^q(5q\\log(s))^n}$$ $$=\\left( \\frac{M}{(2s)^{1/n}q^{1/q}(5\\log(s))^{1/q}} \\right)^{qn}=M_{r+1}^{q_{r+1}}$$ vectors of length $qn=q_{r+1}$.  \nRecall that $n=\\frac{100}{\\epsilon}(\\log(s))^2$ and $q_0=(1+o(1))100s^2$, \nso when we iterate the Markov chain argument $\\log_2(s)$ times, we lose a factor of at most \n$$\n\\frac{M_0}{M_{\\log_2(s)}} \\le (2s)^{2\\log(s)/n}\n\\left(\\prod_{j=0}^{\\infty}(q_0n^j)^{1/q_0n^j}\\right)(5\\log(s))^{1/q_0}\n(5\\log(s))^{2\\log(s)/n}\n$$ \n$$\n\\le (10s\\log(s))^{2\\log(s)/n}2^{10\\log(s)/s^2}(5\\log(s))^{1/100s^2}.\n$$ \nIn the last inequality here\nthe bound on the second term holds because the relevant infinite product  \nis at most $(q_0^{1/q_0})^{100}<2^{10 \\log s/s^2}$.\nFor large $s$ this is easily seen to be smaller than, say,\n$1+\\epsilon/4$. Since for large $s$, $M_0>2-\\epsilon/2$, for \n$s>s_0(\\epsilon)$ we get $M_{\\log_2(s)} >2-\\epsilon$.\nThus at the end we have \n$q=100s^2\\left(\\frac{O(1)}{\\epsilon}(\\log(s))^2\\right)^{\\log_2(s)}$ \nand at least $(2-\\epsilon)^q$ vectors of length $q$ which cover $Z_s$.  \nOne can easily modify the argument to work for any larger $q$, or simply \nuse the super-multiplicative property of $R(s,q)$ (see the \nbeginning of Section 4)\nto conclude, \ntaking $\\epsilon=1/\\log s$,  that for every large $s$ and for all\n$q>s^{(3+o(1)) \\log \\log s}$, $R(s,q) > (2-\\frac{1}{\\log s})^q$.\nThis completes the proof of Theorem \\ref{t13}, part 2, for prime $s$.\n\n\\subsection{General $s$}\n\nGiven an arbitrary fixed integer $s>2$, we now show how to find $(2-o(1))^q$ vectors in $Z_s^q$ which form a $Z_s$-covering family; this shows \n$R(s,q) \\ge (2-o(1))^q$ even for composite $s$.  \nThe general strategy is similar to our strategy in the previous section, except now we work over $Z$, and the set we are covering does not grow \nso quickly. As before, we are not concerned with the vectors \ncovering $0$, because we can simply add an extra \ncoordinate to deal with it.   Since the argument is very similar to\nthe one described in the previous subsection, we only provide a brief\ndescription omitting some of the formal details.\n\n\nWe prove the stronger statement that for any fixed $s$, we can find a $(2-o_q(1))^q$ size family over $Z$ that covers $[-s,s]$, i.e. the difference of any two vectors contains all integers between $-s$ and $s$ as coordinates.  Let $S$ be the least common multiple of the first $s$ positive integers.  We will assume without loss of generality that $2S \\mid q$. \n\nAt the first step of our iteration, we consider vectors over $Z^q$ \nwith an equal number of each element of $[2S]$ as coordinates so that the first $\\frac{q}{S}$ coordinates are $1$s and $2$s in some order, \nthe next $\\frac{q}{S}$ are an equal number of $3$s and $4$s, and so on.  \nWe can find $(2-o_q(1))^q$ of these and they cover $\\{ \\pm 1\\}$.  Furthermore, they have the property that for any ordered pair of these vectors, there is a coordinate in which the first has an even integer $2k$ and the second has the odd integer $2k-1$ for some $1 \\le k \\le S$.  This property is crucial and maintained throughout our iterations.\n\nNow we describe the $m$th step of our iteration, for $2 \\le m \\le s$.  Define a bijection $f=f_m$ on $[2S]$ so that for all integers $1 \\le k \\le S$, $f(2k)=f(2k-1)+m$.  \nWe can do this for instance by setting $f(1)=1, f(2)=m+1, f(3)=2, f(4)=m+2, \\cdots, f(2m-1)=m, f(2m)=2m$ and then set $f(x)=f(x-2m)+2m$ for $x>2m$ as long as $x \\le 2S$.\n\nWe then apply the same Markov chain argument as before, defining \nsets $S_i$ and constructing words $x_1x_2 \\cdots$ \nstarting and ending at the same vectors.  \nNow, however, when $x_i$ is in some set $S_j$, instead of \ndemanding $x_{i+1} \\in \\alpha S_j$ we require\n$x_{i+1} \\in f(S_j)$, meaning that if we apply $f^{-1}$ to each coordinate of $x_{i+1}$, the result will be in $S_j$.  \nLooking at the first place where two vectors differ gives us a difference of $\\pm m$.  Looking at the last place where they differ shows that the crucial property is preserved, and also that the differences $\\pm 1, \\cdots, \\pm (m-1)$ are retained.  \n\nThus after $s=O(1)$ iterations, this algorithm produces $(2-o_q(1))^q$ \nvectors of length $q$ which cover all of $[-s,s]$, after we add \nan extra coordinate to deal with covering $0$. This completes the proof\nof Theorem \\ref{t13}. \n\n\\section{Concluding remarks and open problems} \n\nA natural open problem is to study the functions \n$R(s,q)$ and $Q(s,r)$ in general.  There are several\nsimple properties that $R(s,q)$ satisfies.  We know that $R(s,q)$ is \n(weakly) increasing in $q$, because to create $r$ covering \nvectors for $Z_s$ of length $q'>q$, we can take $r$ vectors for $Z_s$ of length $q$ and pad them with $q'-q$ zeroes at the end.  \nFurthermore, we know that $R(s,q)$ is super-multiplicative, i.e. $R(s,q_1+q_2) \\ge R(s,q_1)R(s,q_2)$, \nbecause if $m=R(s,q_1)$, $n=R(s,q_2)$ then we can find \nvectors $v_1, \\cdots, v_m$ of length $q_1$ and $w_1, \\cdots, w_n$ of \nlength $q_2$ that form covering families. The $mn$ vectors $v_iw_j$ \nobtained by concatenating $v_i$ and $w_j$ \nare clearly a covering family of length $q_1+q_2$.\n\nFor $q<s$, we have $R(s,q)=1$, because of course we can take a single vector in $Z_s^q$, but if we take two then their difference can only \ncover a set of size $q$ and cannot cover $Z_s$.\n\nThe next natural question is studying the value of $R(s,s)$.\n\n\\begin{prop}\n\\label{p41}\n$R(s,s) \\le s$, and $R(s,s) \\ge p$ where $p$ is the smallest prime \nfactor of $s$.  When $p=2$ this is tight, that is,\nif $s$ is even, then $R(s,s)=2$.\n\\end{prop}  \n\n\\begin{proof}\n\tFor the lower bound, for each $0 \\le a<p$ we have a vector $(0,a,2a, \\cdots, (s-1)a)$ reduced modulo $s$.  This is a covering system for $Z_s$, since all positive integers smaller than $p$ are relatively prime to $s$.\n\t\n\tFor the upper bound, assume there was a covering family in $Z_s^s$ with $s+1$ vectors, so that the difference of any two of these vectors has all values of $Z_s$ exactly once.  \nBy the pigeonhole principle, there exist \ntwo vectors $v$ and $w$ so that the difference of \ntheir first and second coordinates is the same.  But then $v-w$ has the same value in its first and second coordinate, a contradiction.\n\t\n\tWhen $s$ is even, $R(s,s) \\ge 2$ as $2$ is the least prime factor of $s$.  If $R(s,s) \\ge 3$ then there exist $3$ \nvectors in $\\mathbb{Z}_s^s$, $(a_1, \\cdots, a_s), (b_1, \\cdots, b_s), (c_1, \\cdots, c_s)$, which are covering.  \nBut then $\\sum (a_i-b_i) \\equiv \\sum_{j=0}^{s-1}j \\equiv (s-1)\\frac{s}{2} \\equiv \\frac{s}{2}$ modulo $s$, and similarly $\\sum (b_i-c_i) \\equiv \\sum (c_i-a_i) \\equiv \\frac{s}{2}$ modulo $s$.  Hence, these three sums are all $\\frac{s}{2}$ modulo $s$, so they must add up to $\\frac{3s}{2} \\equiv \\frac{s}{2}$ modulo $s$.  But they add up to $0$, so this is a contradiction.\n\\end{proof}\n\nNonetheless, the problem of determining $R(s,s)$ for all $s$ \nremains open, and the lower bound in the last proposition\nis not tight.  \nA computer search gives that $R(15,15) \\ge 4$ (\\cite{L}).  One example is the $4$ vectors given by the rows of the matrix $$\\begin{pmatrix} \n0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\\\ \n0 & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 \\\\  \n0 & 2 & 1 & 5 & 7 & 9 & 12 & 14 & 13 & 3 & 6 & 4 & 10 & 8 & 11 \\\\\n0 & 3 & 9 & 1 & 10 & 14 & 7 & 11 & 4 & 12 & 5 & 8 & 2 & 6 & 13 \\end{pmatrix} $$ \nwhich are covering for $Z_{15}$.\n\nFor $q$ which is only a little bigger \nthan $s$, we can prove a reasonable upper bound, \nusing essentially the same observation applied in the \nproof that $R(s,s) \\le s$.\n\n\\begin{prop}\n\tIf $2(q-s)^2<s-1$, then $R(s,q) \\le s+2(q-s)+2(q-s)^2$. \n\\end{prop}\n\n\\begin{proof}\n\tAssume a contradiction for some $q>s$.  \nBy the definition of $R(s,q)$, for any $r \\le R(s,q)$ there exists a matrix of $r$ rows and $q$ columns so that the difference of any two rows contains all values modulo $s$; in particular there exists such a matrix for $r=s+1+2(q-s)+2(q-s)^2$.  \nWe will double count the number of pairs of rows and columns $r_i, r_j, c_k, c_{\\ell}$ with the following property: the $2 \\times 2$ submatrix formed by $r_i, r_j, c_k, c_{\\ell}$ has sums of opposite corners equal.\n\t\n\tFor any two rows $r_i, r_j$, we examine their difference.  This is a vector over $Z_s$ with $q$ entries containing all elements modulo $s$, so at most $\\dbinom{q-s+1}{2}$ pairs of its entries can be equal.  \nThis means that for the pair of rows $r_i, r_j$, there exists at most $\\dbinom{q-s+1}{2}$ pairs of columns $c_k, c_{\\ell}$ that satisfy our property.  \nSo the total number of such pairs as $1 \\le i,j \\le r$, $1 \\le k, \\ell \\le q$ range is at most $\\dbinom{r}{2}\\dbinom{q-s+1}{2}$.  However, if we instead fix $c_k, c_{\\ell}$ we see that the difference of these two columns is a vector of length $r$ over $Z_s$ which has at least $r-s$ pairs of equal \nelements. This means that \nthere are at least $\\dbinom{q}{2}(r-s)$ $r_i, r_j, c_k, c_{\\ell}$ with our property.  \nHence \n$$\n\\dbinom{q}{2}(r-s) \\le \\dbinom{r}{2}\\dbinom{q-s+1}{2}\n$$ and so $2q(q-1)(r-s) \\le r(r-1)(q-s+1)(q-s).$  \nBy our assumption, $2q>r$ and so $2(q-1) \\ge r-1$.  Hence $2q(q-1)(r-s) \\le r(r-1)(q-s+1)(q-s) \\le 4q(q-1)(q-s+1)(q-s).$  Thus $r-s \\le 2(q-s+1)(q-s)=2(q-s)+2(q-s)^2$ and so $r \\le s+2(q-s)+2(q-s)^2$, a contradiction.  \\end{proof}\n\t\t\nNote that this argument gives a bound only when $q=s+O(\\sqrt{s})$, \nand the bound is $s+O((q-s)^2)$.  We believe that this is not tight. \n\t\nThere is a nontrivial (though weak) \nbound which holds in the regime $q=s+\\omega(\\sqrt{s})$ as long as $q<Cs$ \nfor some fixed constant $C< \\log_2 e$.  \nThe natural open problem here is to study cases when $q$ is larger but still fairly small, for instance if $q=1.5s$, $q=s\\log(s)$, or $q=s^2$.  Furthermore, it would be interesting to figure out how large $q$ has to be so that $R(s,q)>(2-\\epsilon)^s$ for a small positive constant $\\epsilon$; the following argument shows that if $\\epsilon$ is small enough, then $\\frac{q}{s}$ must be bigger than an absolute constant which is above $1.44$, but we believe this is far from tight.\n\n\\begin{prop} \nIf $1 \\le \\frac{q}{s}<C<\\log_2(e)$, \nthen $R(s,q) \\le (2-\\varepsilon)^q$ for $\\varepsilon=\\varepsilon(C)>0$.\n\\end{prop}\n\n\\begin{proof}\nSay that $R(s,q) \\ge r$ so there are $r$ vectors in $Z_s^q$ which are covering.  Place them in a matrix with $r$ rows and $q$ columns.  Now, for each column select uniformly an element of $Z_s$, and add it to all the elements of that column, reducing modulo $s$.  This does not change the covering property.  \t\nAfter having done this, replace all entries of this matrix by their reduction modulo $2$.  Given two entries of a matrix in the same column and different rows, if initially they differed by $k$ modulo $s$ they now have some probability $p_k$ of being equal that depends only on $k$, \nwhich is the probability that if $x \\in Z_s$ is chosen randomly \nand uniformly \nthen the reductions of $x,x+k$ modulo $s$ have the same parity.  One can easily check that $p_0=1$, $p_1=1/s$, $p_2=(s-2)/s$, and so on, \nso that $\\prod_{i \\in Z_s}p_i=e^{-s(1+o(1))}$ by Stirling's formula.  \nFurthermore, if we take this product over all $i \\in Z_s$ except \nfor a subset of size \n$2cs$ for a small constant $c$, we have a bound of $e^{-s(1-\\epsilon)}$, where $\\epsilon$ is a small positive constant that tends to $0$ with $c$.\n\nLet $K=cs$.  For any pair of rows, the probability their colors match in all but at most $2K$ places is at most \n$e^{-s(1-\\epsilon')}$ for some $\\epsilon'$ that tends to $0$ with $c$, \nby a union bound over all \nthe at most $2\\dbinom{q}{2K}$ \npossible sets of places where they do not match.  Thus if we define\na graph on our $r$ vectors where two are adjacent iff their values \nupon reduction modulo $2$ match in all but at most $2K$ places, \nthis graph will in expectation have edge density $e^{-s(1-\\epsilon')}$ \nand thus for some fixed choice of the random shifts\nwill have at most $r^2e^{-s(1-\\epsilon')}$ edges.  Fixing these shifts\none can remove less than \n$r^2e^{-s(1-\\epsilon')}$ vertices from this graph and get an \nindependent set. But this set\nmust be of size at most $(2-\\delta)^q$ for some $\\delta=\\delta(c) > 0$,\nbecause this size is bounded by the cardinality \nof a family of disjoint Hamming balls in $\\{0,1\\}^q$, each of \nradius $K=cs$.  Thus we have $r \\le r^2e^{-s(1-\\epsilon')}+(2-\\delta)^q$, \nand the same inequality holds for any $r'<r$ by applying the same\nreasoning to a set of $r'$ of our vectors.\n\nNow if $q<Cs$ for some fixed  $C< \\log_2 e$ then setting $q > q_0(\\epsilon',C)$, $r'=3 (2-\\delta)^q$ violates the \nlast inequality (with $r$ replaced by $r'$) since for this value\nof $r'$ and large $q$, $(r')^2 e^{-s(1-\\epsilon')}<r'/2$ and\n$(2-\\delta)^q <r'/2$ so their sum is smaller than $r'$.\nThis establishes the assertion of the proposition.\n\\iffalse\nThis means that if $e^{s(1-\\epsilon')}>(2-\\delta)^q$ then this forces $r \\le \\frac{1}{2}(2-\\delta)^q$, since if $r$ were bigger, we could apply the same argument to a subset of these vectors.  \nIn particular, this bounds $r$ to be at most $(2-\\varepsilon)^q$ for some $\\varepsilon>0$ whenever $e^{s(1-\\epsilon')}>(2-\\delta)^q$.  Since $\\delta, \\epsilon', \\varepsilon \\to 0$ as $c \\to 0$, we can always find some $c$ depending on $q/s$ so that this is true, as long as $\\frac{q}{s}<\\log_2(e)$.  \nThen $\\varepsilon$ depends only on $c$ and thus only \non $\\frac{q}{s}$, so we are done. \n\\fi\n\\end{proof}\n\nOn the opposite extreme, one may ask \nwhat happens if $s \\ge 3$ is a small fixed positive integer \nand $q$ grows.  We know by Proposition~\\ref{p31} that $R(2,q)=2^{q-1}$, \nso it is natural to ask what happens when $2$ is replaced by \na larger positive integer.\n\n\\begin{conj}\nFor any fixed $s \\ge 3$, $R(s,q)=o(2^q)$, \nand furthermore $R(s,q)=\\Theta(2^q/q^{c})$ where \n$c=c(s)>0$ is a constant that depends only on $s$.\n\\end{conj}\n\nNote that the vectors over $Z_3^q$ with a zero in the first coordinate and with exactly $\\lfloor q/2 \\rfloor$ ones and $\\lceil q/2 \\rceil$ zeroes are a covering system, so $R(3,q)=\\Omega(2^q/\\sqrt{q})$.  \nWhen $s$ is odd the best upper bound known for $R(s,q)$ \nis $O(2^{q})$, as shown in Proposition \\ref{p31}.  \nWhen $s$ is even, if the vectors of a covering system for $Z_s^q$ are reduced modulo $2$, any two differ in at least $s/2$ places, \nso there are at most $O(2^q/q^{\\lfloor \\frac{s-1}{4} \\rfloor})$ of them, \nas the Hamming balls of radius $\\lfloor \\frac{s-1}{4} \\rfloor$ \ncentered at these reduced vectors are pairwise disjoint.\nIt would be interesting to establish the above conjecture and to find the relevant constants $c$ if they indeed exist.  \nIn particular it seems plausible that \n$c=1/2$ when $s=3$, and if so then the lower bound for $R(3,q)$ is \nessentially optimal.  We conjecture that this is indeed the case.\n\n\\begin{conj}\n\t$R(3,q)=\\Theta(2^q/\\sqrt{q}).$\n\\end{conj}\n\nIn \\cite{Ca} it is shown that $R(3,q) \\le (\\frac{1}{2}+o(1))2^q$ \nwhen $q$ is even and $R(3,q) \\le (\\frac{1}{3}+o(1))2^q$ when $q$ \nis odd, leaving this problem open.  Note that we do not even know \nhow to prove the weaker claim that a $\\{-1,0,1\\}$ covering system \n(or equivalently just a $\\{\\pm 1\\}$ covering system) \nover $Z$ must have size $o(2^q)$.\n\\vspace{0.2cm}\n\n\\noindent\nOur original motivation for studying the function $R(s,q)$ here is\nits connection to the product dimension $Q(s,r)$ of the disjoint union of\n$r$ cliques, each of size $s$.  This connection is described in Proposition\n\\ref{p12}. The results here suffice to determine the asymptotic behaviour\nof $Q(s,r)$ for every fixed $s$ as $r$ tends to infinity, but do not provide\ntight bounds when $r$ is not much bigger than $s$. \nWe conclude this short paper with several simple comments about this range\nof the parameters. The first remark is that the product dimension\n$Q(s,r)$ is at least $s$ for every $r \\geq 2$. To see this fix a \nvertex $u$ of the first clique and observe that in every proper coloring of\nthe graph $K_s(r)$ of $r$ disjoint cliques, each of size $s$, there is\nat most one pair $uv$ with a vertex $v$ of the second clique so that $u$\nand $v$ have the same color.  As altogether there  are $s$ such pairs, \nand each one has to be monochromatic in at least one vertex coloring \nin a collection exhibiting an upper bound for the product dimension, the\nnumber of such colorings is at least $s$.  Another comment is that\n$Q(s,r)$ is clearly monotone non-decreasing in both $r$ and $s$. Therefore,\nfor every $r,s \\geq 2$, $Q(s,r) \\geq Q(2,r) = \\lceil \\log_2(2r) \\rceil$.\n\nA {\\em transversal design} $TD(r,s)$ of order $s$ and block size $r$ (with \nmultiplicity $\\lambda=1$) is a set $V$ of $sr$ elements partitioned \ninto $r$ pairwise disjoint groups, each of size $s$, and a collection\nof blocks, each containing exactly one element of each group, so that\nevery pair of elements from distinct groups is contained in exactly\none block. A transeversal design is {\\em resolvable}\nif its blocks can be partitioned\ninto parallel classes where the blocks in any parallel class \npartition the set $V$.\nThere is a substantial amount of literature about transversal\ndesigns, see \\cite{CD}. It is not difficult to check that \n$Q(s,r)=s$ if and only if a resolvable \n$TD(r,s)$ exists and hence the known results\nabout resolvable \ntransversal designs supply nearly precise information for the range\n$r \\leq s$ (it is easy to see that such a design cannot exist for \n$r>s$).  In particular, for every prime power $s$, $Q(s,s)=s$\nand therefore  by the obvious monotonicity, for any prime power $s$,\n$Q(s,r)=s$ for every $2 \\leq r \\leq s$, and if $p$ is a prime power then for\nevery $s,r \\leq p$, $Q(s,r) \\leq p$. (Note that \nPropositions \\ref{p12} and \\ref{p41}  also imply that \n$Q(s,s)=s$ when $s$ is\na prime.)\n\nIt is not difficult to prove that for every $s,r_1,r_2$,\n\\begin{equation}\n\\label{e999}\nQ(s,r_1r_2) \\leq Q(s,r_1)+Q(s,r_2). \n\\end{equation}\nIndeed, given the graph\n$K_s(r_1r_2)$ consisting of $r_1r_2$ disjoint cliques, each of size\n$s$, we can split the cliques into $r_1$ disjoint groups, each consisting\nof $r_2$ cliques. Define $Q(s,r_1)$ \nproper colorings in which the cliques\nin every group are colored the same, based on the system of colorings that\nshows that the product dimension of $K_s(r_1)$ is $Q(s,r_1)$. Add to these\n$Q(s,r_2)$ additional colorings, whose restrictions to the $r_2$ cliques\nin each group are exactly the colorings showing that the product dimension\nof $K_s(r_2)$ is $Q(s,r_2)$. The resulting $Q(s,r_1)+Q(s,r_2)$ colorings\nestablish (\\ref{e999}).\n\nThe above comments together with the results in the previous sections \nprovide upper and lower bounds for $Q(s,r)$ for all\n$s$ and $r$, but these bounds are quite far from each other when\n$r$ is much bigger than $s$ but much smaller than $2^{s^{3 \\log \\log s}}$. \nIn particular, for $r=2^s$ the bounds we have are\n$$\ns \\leq Q(s,2^s) \\leq (1+o(1)) \\frac{s^2}{\\log s}.\n$$\nIt would be interesting to close this gap.  \n\\vspace{0.2cm}\n\n\\noindent\n{\\bf Acknowledgment}\nWe thank J\\'anos K\\\"orner, Robert Kleinberg and Matt Weinberg  \nfor helpful comments.\n\n", "meta": {"timestamp": "2019-05-28T02:04:34", "yymm": "1905", "arxiv_id": "1905.10483", "language": "en", "url": "https://arxiv.org/abs/1905.10483", "abstract": "The product dimension of a graph $G$ is the minimum possible number of proper vertex colorings of $G$ so that for every pair $u,v$ of non-adjacent vertices there is at least one coloring in which $u$ and $v$ have the same color. What is the product dimension $Q(s,r)$ of the vertex disjoint union of $r$ cliques, each of size $s$? Lovász, Nešetřil and Pultr proved in 1980 that for $s=2$ it is $(1+o(1)) \\log_2 r$ and raised the problem of estimating this function for larger values of $s$. We show that for every fixed $s$, the answer is still $(1+o(1)) \\log_2 r$ where the $o(1)$ term tends to $0$ as $r$ tends to infinity, but the problem of determining the asymptotic behavior of $Q(s,r)$ when $s$ and $r$ grow together remains open. The proof combines linear algebraic tools with the method of Gargano, Körner, and Vaccaro on Sperner capacities of directed graphs.", "subjects": "Combinatorics (math.CO)", "title": "On the product dimension of clique factors", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9891815491146485, "lm_q2_score": 0.8558511506439708, "lm_q1q2_score": 0.8465921670055574}}
{"url": "https://arxiv.org/abs/1801.00517", "title": "Dedekind Sums with Even Denominators", "abstract": "Let $S(a,b)$ denote the normalized Dedekind sum. We study the range of possible values for $S(a,b)=\\frac{k}{q}$ with $\\gcd(k,q)=1$. Girstmair proved local restrictions on $k$ depending on $q\\pmod{12}$ and whether $q$ is a square and conjectured that these are the only restrictions possible. We verify the conjecture in the cases $q$ even, $q$ a square divisible by $3$ or $5$, and $2\\le q\\le 200$ (the latter by computer), and provide progress towards a general approach.", "text": "\\section{Introduction}\n\n\\begin{defn}\nFor coprime integers $a$ and $b$ with $b>0$,  the \\textit{Dedekind sum} $s(a,b)$ is defined as\n\\[\ns(a,b) = \\sum_{k=1}^{b} \\left(\\!\\!\\left(\\frac{k}{b}\\right)\\!\\!\\right)\\left(\\!\\!\\left(\\frac{ak}{b}\\right)\\!\\!\\right),\n\\]\nwhere $ {\\displaystyle (\\!(\\,)\\!):\\mathbb {R} \\rightarrow \\mathbb {R} }$ denotes the \\textit{sawtooth function}, defined by\n\\[\n(\\!(x)\\!) = \\begin{cases}\n\\{x\\}-\\frac{1}{2}&x \\not \\in \\Z\\\\\n0 &x \\in \\Z.\n\\end{cases}\n\\]\nWe will primarily work with the \\textit{normalized Dedekind sum} $S(a,b)$, defined by \n\\[\nS(a,b)=12s(a,b),\n\\] \nwhich will make computation more convenient.\n\n\\end{defn}\n\nThe definition of the Dedekind sum is motivated by its use in the transformation law of the Dedekind eta function (\\cite[p. 52]{Apostol76}). Dedekind sums have been studied in a variety of contexts, including in applications to algebraic geometry, lattice point enumeration, and the study of modular forms (\\cite{Urzua07,Rademacher72,Apostol76,Bruggeman94}). The distribution of possible values of Dedekind sums has also been considered extensively (\\cite{Vardi93,Bruggeman94,Hickerson97,Myerson98,Girstmair15,GirstmairIntegers2017,GirstmairLargest2017,GirstmairEqual2016}). \n\nIt was noted by Rademacher and Grosswald in \\cite[p. 28]{Rademacher72} that the range of values of $S(a,b)$ is unknown, which is our central question. Hickerson (\\cite{Hickerson97}) proved that this range is dense in $\\R$, and Girstmair (\\cite{Girstmair15}) found that each rational number $r \\in [0,1)$ occurs as the fractional part of a Dedekind sum $S(a,b)$. Furthermore, Girstmair (\\cite{GirstmairNovember2017}) proved that each value in this range occurs as a Dedekind sum infinitely many times in a nontrivial sense. Recently, Girstmair (\\cite{Girstmair17}) classified the denominator of a Dedekind sum in terms of $a$ and $b$: if $S(a,b) = \\frac{k}{q}$ with $\\gcd(k,q)=1$, then $q = \\frac{b}{\\gcd(a^2+1,b)}$. Girstmair also conjectured that for a fixed integer $q\\ge 2$, the set of possible $k$ coprime with $q$ such that $\\frac{k}{q} = S(a,b)$ for some $a,b$ are exactly the integers $k$ for which \n\\begin{itemize}\n\\item If $3\\nmid q$, then $3\\mid k$.\n\\item If $2\\nmid q$, then\n\\[\nk \\equiv \\begin{cases}\n2\\pmod{4} & q\\equiv 3 \\pmod{4}\\\\\n0 \\pmod{8} & q\\text{ is a square}\\\\\n0\\pmod{4}&\\text{else.}\n\\end{cases}\n\\]\n\\end{itemize}\nand showed that these conditions are indeed necessary. Finally, Girstmair proved that if $k\\equiv k' \\pmod{q(q^2-1)}$, then $\\frac{k}{q}$ is the value of a normalized Dedekind sum if and only if $\\frac{k'}{q}$ is the value of a normalized Dedekind sum, effectively reducing the problem to a finite existence problem $\\pmod{q(q^2-1)}$ for each $q$. In particular, this reduction allowed Girstmair to verify the conjecture for all $q\\le 60$ by computer.\n\nIn Section \\ref{sec:reduce} of the present paper, we establish a decomposition $qS(a',b') = \\lambda(a,t,t^{*}) + \\Delta(at^{*}+j)$ for certain numerators of normalized Dedekind sums and analyze $\\lambda$ and $\\Delta$ $\\pmod{q}$ and $\\pmod{q^2-1}$ to reduce Girstmair's conjecture to an existence problem for $\\lambda \\pmod{q^2-1}$. In Section \\ref{sec:partial}, we specialize $\\lambda\\pmod{q^2-1}$ to a function $f(a)$, which we then split into a linear and a periodic part based on a generalization of Rademacher's three-term relation used by Girstmair. If the slope of the linear part is small enough (in particular if $\\gcd(a,q)$ is small enough), this observation is enough to prove Girstmair's conjecture. In particular, we prove the conjecture holds for all $q$ even and all perfect squares $q$ which are divisible by $3$ or $5$. We also use computer verification and our first observation to prove the conjecture for all $2\\le q\\le 200$.\n\nAs noted in \\cite{Girstmair17}, the case $q=1$ has been resolved completely, so for the remainder of the paper we assume $q\\ge 2$.\n\n\\section{Background}\n\n\nNote that given a fixed $b$, the normalized Dedekind sum only depends on the residue class of $a\\pmod{b}$.\n\nA classical fact about Dedekind sums reveals unexpected symmetry:\n\n\\begin{lemma}[Reciprocity Law]\n\\label{recip}\nIf $a$ and $b$ are coprime positive integers, then\n\\[\nS(a,b) +S(b,a) = \\frac{a}{b}+\\frac{b}{a}+\\frac{1}{ab}-3.\n\\]\n\\end{lemma}\n\\begin{proof}\nSee, for example, \\cite[p. 27]{Rademacher72}.\n\\end{proof}\nThe reciprocity law is crucial to the study of Dedekind sums. As an example, along with the fact that $S(a-nb,b) = S(a,b)$ for $n \\in \\Z$, the reciprocity law yields an alternative method for computing $S(a,b)$ by following the Euclidean algorithm.\n\n\nGiven coprime integers $a$ and $b$ with $b>0$, it is natural to ask what the denominator of $S(a,b)$ is. It follows from expansion of the original definition and some rearrangement (see, for example, \\cite[p. 27]{Rademacher72}) that $bS(a,b) \\in \\Z$, and so the denominator of $S(a,b)$ when written as a reduced fraction is a divisor of $b$. In fact, a more exact statement holds:\n\\begin{thm}[Girstmair, \\cite{Girstmair17}]\n\\label{denom}\nSuppose $a$ and $b$ are coprime integers with $b>0$, and suppose we can write $S(a,b)$ as\n\\[\nS(a,b) = \\frac{k}{q}\n\\]\nfor $k,q \\in \\Z$, $q>0$, and $\\gcd(k,q)=1$. Then\n\\[\nq = \\frac{b}{\\gcd(b,a^2+1)}.\n\\]\nIn particular, given positive integers $b,q$ and an integer $a$ with $\\gcd(a,b)=1$, $S(a,b)$ has the form $\\frac{k}{q}$ for some $k\\in \\Z$ with $\\gcd(k,q)=1$ if and only if\n\\[\nb = \\frac{q(a^2+1)}{t}\n\\]\nfor some positive integer $t$ with $\\gcd(t,q)=1$.\n\\end{thm}\nFurthermore, Girstmair proved that the problem of classifying the set of $k\\in \\Z$ such that $\\gcd(k,q)=1$ and $\\frac{k}{q}$ is the value of a normalized Dedekind sum can be reduced to classification $\\pmod{q(q^2-1)}$. More precisely, he proved the following.\n\n\\begin{thm}[Girstmair, \\cite{Girstmair17}]\n\\label{finitemod}\nLet $k$ and $q$ be coprime integers with $q\\ge 2$. If $k' \\in \\Z$ and \n\\[\nk' \\equiv k \\pmod{q(q^2-1)},\n\\]\nthen $\\frac{k}{q}$ is the value of a normalized Dedekind sum if and only if $\\frac{k'}{q}$ is the value of a normalized Dedekind sum.\n\\end{thm}\nThus for each fixed value of $q$, the question of determining the set of normalized Dedekind sums with denominator $q$ is reduced to determining the finite set of residue classes $\\pmod{q(q^2-1)}$ which represent the possible numerators. In fact, Girstmair provides necessary conditions on the numerators of such normalized Dedekind sums and conjectures that these are the only restrictions on the set of normalized Dedekind sums with a given denominator.\n\\begin{thm}[Girstmair, \\cite{Girstmair17}]\n\\label{qgood}\nSuppose for integers $a,b,k,q$ with $b\\ge 1,q\\ge 2,$ and $\\gcd(a,b)=\\gcd(k,q)=1$, it holds that $S(a,b) = \\frac{k}{q}$. Then\n\\begin{itemize}\n\\item If $3\\nmid q$, then $3\\mid k$.\n\\item If $2\\nmid q$, then\n\\[\nk\\equiv\\begin{cases}\n2\\pmod{4} & q\\equiv 3 \\pmod{4}\\\\\n0\\pmod{8} & q \\text{ is a square}\\\\\n0 \\pmod{4} & \\text{else.}\n\\end{cases}\n\\]\n\\end{itemize}\n\\end{thm}\n\n\\begin{conj}[Girstmair, \\cite{Girstmair17}]\n\\label{conj1}\nFor integers $k,q$ with $q\\ge 2$ and $\\gcd(k,q)=1$, there exist integers $a,b$ with $b\\ge 1$ and $\\gcd(a,b)=1$ such that\n\\[\nS(a,b) = \\frac{k}{q}\n\\]\nif and only if the conditions of Theorem \\ref{qgood} hold.\n\\end{conj}\nThis conjecture is our main focus of study.\n\n\n\\section{Reduction to \\texorpdfstring{$\\pmod{q^2-1}$}{mod}}\n\\label{sec:reduce}\nWe first reduce the problem from a question of existence $\\pmod{q(q^2-1)}$ to an existence problem $\\pmod{(q^2-1)}$. In proving Theorem \\ref{finitemod}, Girstmair uses the generalized three-term relation (see \\cite{Girstmair98}) to obtain a new formula for the normalized Dedekind sum in light of Theorem \\ref{denom}.\n\n\\begin{lemma}[Girstmair, \\cite{Girstmair17}]\n\\label{girstmairidentity}\nLet $q,t$ be positive integers such that $q\\ge 2$ and $\\gcd(t,q)=1$, let $a$ be an integer such that $\\gcd(a,q)=1$ and $t\\mid a^2+1$, and let $b=\\frac{q(a^2+1)}{t}$. Then\n\\[\nS(a,b) = \\frac{(q^2-1)a}{tq}-S(aq,t)+S(at^{*},q),\n\\]\nwhere $t^{*}$ is any integer such that $t t^{*}\\equiv 1 \\pmod{q}$.\n\\end{lemma}\nNow fix an integer $a$, no longer necessarily coprime with $q$, such that $t\\mid a^2+1$. We consider Dedekind sums in the form $S(a',b')$ where $a'\\equiv a \\pmod{t}$ and $\\gcd(a', q)=1$, as motivated by \\cite{Girstmair17}. If $a'= a+tj$ for some $j$ such that $\\gcd(a',q)=1$ and $b' = \\frac{q(a'^2+1)}{t}$, then $t\\mid a'^2+1$ still holds, so $S(a',b')$ has reduced denominator $q$. Its numerator is\n\\begin{align*}\nq S(a',b') &=\\frac{(q^2-1)a}{t} +(q^2-1)j - qS(a'q,t)+qS(a't^{*},q)\\\\\n&=\\frac{(q^2-1)a}{t} +(q^2-1)j - qS(aq,t)+qS(at^{*}+j,q)\\\\\n\\begin{split}&= \\left(\\frac{(q^2-1)a}{t}-(q^2-1)at^{*}-qS(aq,t)\\right)\\\\\n&\\quad +\\left((q^2-1)(at^{*}+j)+q S(at^{*}+j,q)\\right).\n\\end{split}\n\\end{align*}\nThe first expression is only dependent on $a,t,$ and $t^{*}$, while the second is only dependent on $at^{*}+j$. This motivates analyzing the two expressions separately. By considering a fixed $a$ and all $a'$ in the form $a'=a+tj$ for varying $j$, we can isolate the behavior of $qS(a',b')$ dependent on $a' \\pmod{t}$ in the first term and the behavior dependent on $a'\\pmod{q}$ in the second term.\n\\begin{defn}\nGiven a fixed integer $q\\ge 2$, we define the functions $\\lambda(a,t,t^{*})$ and $\\Delta(\\ell)$ by\n\\[\n\\lambda(a,t,t^{*}) = \\frac{(q^2-1)a}{t} - (q^2-1)at^{*} - qS(aq,t)\n\\]\nand\n\\[\n\\Delta(\\ell) = (q^2-1)\\ell+qS(\\ell,q)\n\\]\nwhere $\\lambda$ is defined on all triples of integers $a,t,t^{*}$ such that $t>0$, $t\\mid a^2+1$, and $tt^{*}\\equiv 1\\pmod{q}$, and $\\Delta$ is defined on all integers $\\ell$ such that $\\gcd(\\ell,q)=1$.\n\\end{defn}\nNote that $\\gcd(a+tj,q)=1$ if and only if $\\gcd(at^{*}+j,q)=1$, so $\\Delta(at^{*}+j)$ is well defined if and only if $\\gcd(a+tj,q)=1$. We can now give our central decomposition of the numerator of $S(a',b').$\n\\begin{prop}\n\\label{split}\nIf integers $a,q,t,j,a',$ and $b'$ are given such that $t$ is positive, $q\\ge 2$, $\\gcd(t,q)=1$, $a'=a+tj$, $b' = \\frac{q(a'^2+1)}{t}$, and $\\gcd(a',q)=1$, then $S(a',b')$ has reduced denominator $q$ and numerator\n\\[\nqS(a',b') = \\lambda(a,t,t^{*}) +\\Delta(at^{*}+j).\n\\]\nFurthermore, $\\lambda(a,t,t^{*})\\in \\Z$ and $\\Delta(at^{*}+j)\\in \\Z$.\n\\end{prop}\n\\begin{proof}\nWe've shown that the identity for $qS(a',b')$ holds and that the denominator of $S(a',b')$ is reduced, so it suffices to show that $\\lambda(a,t,t^{*})\\in \\Z$ and $\\Delta(at^{*}+j) \\in \\Z$. But the denominator of $S(\\ell,q)$ divides $q$, which implies that $qS(\\ell,q)$ and therefore $\\Delta(at^{*}+j)$ are integers. Furthermore $qS(a',b')\\in \\Z$, which implies $\\lambda(a,t,t^{*}) = qS(a',b')-\\Delta(at^{*}+j) \\in \\Z$.\n\\end{proof}\nTo understand the possible values of $qS(a',b')\\pmod{q(q^2-1)}$, we analyze the behavior of $\\lambda(a,t,t^{*})$ and $\\Delta(\\ell)$ when reduced $\\pmod{q}$ and $\\pmod{q^2-1}$. \n\\begin{lemma}\n\\label{deltamodq}\nIf $\\ell_1,\\ell_2\\in \\Z$ such that $\\gcd(\\ell_1,q)=\\gcd(\\ell_2,q)=1$ and $\\ell_1\\equiv\\ell_2\\pmod{q}$, then $\\Delta(\\ell_1)\\equiv \\Delta(\\ell_2)\\pmod{q}$. Furthermore, as $\\ell\\pmod{q}$ ranges over $(\\Z/q\\Z)^{\\times}$, $\\Delta(\\ell)\\pmod{q}$ ranges over $(\\Z/q\\Z)^{\\times}$. In other words, $\\Delta$ is a bijection when considered as a mapping from $(\\Z/q\\Z)^{\\times}$ to itself.\n\n\\end{lemma}\n\\begin{proof}\nThe first assertion is clear from the periodicity of $S(\\ell,q)$ in $\\ell$. Now we claim\n\\[\nq S(\\ell,q) \\equiv \\ell+\\ell^{*} \\pmod{q}\n\\]\nwhere $\\ell^{*}$ is an integer such that $\\ell \\ell^{*} \\equiv 1\\pmod{q}$. Indeed, if $c$ is an integer such that $\\ell\\ell^{*}-cq=1$, and we define\n\\[\nM = \\begin{pmatrix}\n\\ell^{*} & c\\\\\nq & \\ell\n\\end{pmatrix} \\in \\SL_2(\\Z),\n\\]\nthen (as $q>0$) the Rademacher function $\\Phi:\\SL_2(\\Z)\\to \\Z$ (see, for example, \\cite[p. 50]{Rademacher72}) is given by\n\\[\n\\Phi(M) = \\frac{\\ell+\\ell^{*}}{q} - S(\\ell,q).\n\\]\nBut since $\\Phi(M) \\in \\Z$, we have $qS(\\ell,q) \\equiv \\ell+\\ell^{*} \\pmod{q}$.\n\nNow reducing $\\Delta(\\ell)\\pmod{q}$, we get\n\\begin{align*}\n\\Delta(\\ell) &\\equiv (q^2-1)\\ell+qS(\\ell,q) \\\\\n& \\equiv -\\ell + (\\ell+\\ell^{*})\\\\\n&\\equiv \\ell^{*} \\pmod{q}\n\\end{align*}\nimplying $\\Delta(\\ell)$ takes the value of exactly the invertible residues $\\pmod{q}$ as $\\ell$ ranges over the invertible residues $\\pmod{q}$.\n\n\\end{proof}\nNow we consider $\\lambda(a,t,t^{*})\\pmod{q}$.\n\\begin{lemma}\n\\label{lambdamodq}\nFor any integers $a,t,t^{*}$ with $t>0$, $t\\mid a^2+1$, $\\gcd(aq,t)=1$, and $tt^{*}\\equiv 1 \\pmod{q}$, we have\n\\[\n\\lambda(a,t,t^{*}) \\equiv 0 \\pmod{q}.\n\\]\n\\end{lemma}\n\\begin{proof}\nNote that $\\gcd(q,t)=1$. Then\n\\begin{align*}\nt \\lambda(a,t,t^{*})&\\equiv (q^2-1)a(1-tt^{*}) - qtS(aq,t)\\\\\n&\\equiv 0\\pmod{q}\n\\end{align*}\nbecause $tt^{*}\\equiv 1\\pmod{q}$ and $tS(aq,t)\\in \\Z$.\n\\end{proof}\nFinally, we consider $\\lambda(a,t,t^{*})\\pmod{q^2-1}$.\n \n\\begin{defn}\nDefine $\\Lambda_q \\subseteq \\Z/(q^2-1)\\Z$ by\n\\[\n\\Lambda_q = \\left\\{\\frac{(q^2-1)a}{t} - qS(aq,t)\\pmod{q^2-1}:\\gcd(aq,t)=1, t \\mid a^2+1,t>0\\right\\}.\n\\]\n\\end{defn}\n\\begin{rem}\nAlternatively, $\\Lambda_q$ gives the range of $\\lambda$:\n\\[\n\\Lambda_q = \\left\\{\\lambda(a,t,t^{*})\\pmod{q^2-1} : \\gcd(aq,t)=1,t\\mid a^2+1,tt^{*}\\equiv 1 \\pmod{q},t>0\\right\\}.\n\\]\n\\end{rem}\nThe characterizations are equivalent because the $(q^2-1)at^{*}$ term vanishes $\\pmod{q^2-1}$.\n\n\\begin{defn}\nDefine $\\Gamma_q\\subseteq \\Z/(q^2-1)\\Z$ by \n\\[\n\\Gamma_q = \\begin{cases}\n\\{24s: s \\in \\Z/(q^2-1)\\Z\\}& q\\text{ is a square}\\\\\n\\{12s: s \\in \\Z/(q^2-1)\\Z\\}&\\text{else.}\n\\end{cases}\n\\]\n\\end{defn}\n\\begin{lemma}\nFor any integer $q\\ge 2$,\n\\[\n\\Lambda_q\\subseteq\\Gamma_q\n\\]\n\\end{lemma}\n\\begin{proof}\nIt suffices to show the following claims:\n\\begin{itemize}\n\\item If $3\\nmid q$, then $3\\mid \\lambda(a,t,t^{*})$.\n\\item If $2\\nmid q$, then\n\\[\n\\lambda(a,t,t^{*})\\equiv \\begin{cases}\n0 \\pmod{8}& q\\text{ is a square}\\\\\n0 \\pmod{4}&\\text{else.}\n\\end{cases}\n\\]\nTo prove both, we note that for any $a$ such that $t\\mid a^2+1$, there exists some $a' = a+tj$ such that $\\gcd(at^{*}+j,q)=1$ and therefore $\\gcd(a',q)=1$, recalling that $\\gcd(t,q)=1$. Then by Proposition \\ref{split}, we may write\n\\[\n\\lambda(a,t,t^{*}) = qS(a',b') - \\Delta(at^{*}+j)\n\\]\nfor $b'=\\displaystyle\\frac{(a'^2+1)q}{t}$.\n\nFirst, suppose $3\\nmid q$. By Theorem $\\ref{qgood}$, $3\\mid qS(a',b')$, and similarly $3 \\mid q S(at^{*}+j,q)$, as the denominator of $S(at^{*}+j,q)$ is a divisor of $q$, which is coprime with $3$. But $3\\mid (q^2-1)$, so $3 \\mid \\Delta(at^{*}+j)$ and $3 \\mid \\lambda(a,t,t^{*})$.\n\nNow suppose $2\\nmid q$. By Theorem \\ref{qgood},\n\\[\nqS(a',b') \\equiv q-1 \\pmod{4}.\n\\]\nFurthermore, $8 \\mid q^2-1$, and by \\cite[p. 34]{Rademacher72},\n\\[\nqS(at^{*}+j,q)\\equiv q+1-2\\left(\\frac{at^{*}+j}{q}\\right)\\pmod{8},\n\\]\nwhere $\\left(\\frac{at^{*}+j}{q}\\right)$ denotes the Jacobi symbol. This implies\n\\[\n\\Delta(at^{*}+j)\\equiv qS(at^{*}+j,q)\\equiv q+1-2\\left(\\frac{at^{*}+j}{q}\\right)\\pmod{8}\n\\]\nand\n\\begin{align*}\n\\lambda(a,t,t^{*})&\\equiv qS(a',b')-\\Delta(at^{*}+j) \\\\\n&\\equiv \\left(q-1\\right)-\\left(q+1-2\\left(\\frac{at^{*}+j}{q}\\right)\\right)\\\\\n&\\equiv -2+2\\left(\\frac{at^{*}+j}{q}\\right)\\\\\n&\\equiv 0\\pmod{4}.\n\\end{align*}\n\\end{itemize}\nFurthermore, if $q$ is a square, then by Theorem \\ref{qgood}, $qS(a',b')\\equiv 0 \\pmod{8}$, and since the Jacobi symbol is multiplicative in the denominator, $\\left(\\frac{at^{*}+j}{q}\\right) \\equiv 1 \\pmod{8}$. Thus\n\\begin{align*}\nqS(at^{*}+j,q) &\\equiv q+1-2\\left(\\frac{at^{*}+j}{q}\\right) \\equiv 1+1-2\\left( 1 \\right) \\equiv 0\\pmod{8}\n\\end{align*}\nimplying $\\lambda(a,t,t^{*})\\equiv qS(a',b') - \\Delta(at^{*}+j)\\equiv 0\\pmod{8}$, as desired.\n\n\\end{proof}\n\nFinally, we present our main conjecture, which specifies the range of $\\lambda(a,t,t^{*})$ and, as we will see, implies Conjecture \\ref{conj1}.\n\\begin{conj}\n\\label{conj2}\nFor any integer $q\\ge 2$,\n\\[\n\\Lambda_q=\\Gamma_q\n\\]\n\\end{conj}\nIn a sense, Conjecture \\ref{conj2} states that $\\lambda(a,t,t^{*})$ takes on all possible values $\\pmod{q^2-1}$ after accounting for the $\\pmod{3}$ and $\\pmod{8}$ restrictions of $\\Gamma_q$. Similarly, this implies that the normalized Dedekind sum takes on all possible values after accounting for the local restrictions given by Theorem \\ref{qgood}.\n\\begin{thm}\nFor each $q$, Conjecture \\ref{conj2} implies Conjecture \\ref{conj1}.\n\\end{thm}\n\\begin{proof}\nSuppose $a,q,t,a',b',$ and $j$ are as in Proposition \\ref{split}. We have established the following facts:\n\\begin{itemize}\n\\item The denominator of $S(a',b')$ is $q$, so its numerator is\n\\[\nqS(a',b') = \\lambda(a,t,t^{*}) + \\Delta(at^{*}+j).\n\\]\n\\item\nIt always holds that\n\\[\n\\lambda(a,t,t^{*})\\equiv 0 \\pmod{q}.\n\\]\n\\item\nAs $at^{*}+j$ ranges over all invertible residues $\\pmod{q}$, the remainder of $\\Delta(at^{*}+j)$ also ranges over all invertible residues $\\pmod{q}$.\n\\item\nAssuming Conjecture \\ref{conj2}, the range of $\\lambda(a,t,t^{*})$ when reduced $\\pmod{q^2-1}$ is $\\Lambda_q=\\Gamma_q$.\n\\end{itemize}\nFor the rest of the proof, assume Conjecture \\ref{conj2} does hold. It is at this point that we finally use our full flexibility to generate pairs $a',b'$ based on a choice of $a$ and a variable choice of $j$. Given any $a,t,$ and $t^{*}$, we may choose $j$ independently to yield new values $a'$ and $b'$ as long as $\\gcd(a',q)=\\gcd(at^{*}+j,q)=1$. Then for any residue $\\lambda(a,t,t^{*})\\pmod{q(q^2-1)}$ and any residue $\\Delta(\\ell)\\pmod{q(q^2-1)}$, the sum $\\lambda(a,t,t^{*}) +\\Delta(\\ell)\\pmod{q(q^2-1)}$ is the residue $\\pmod{q(q^2-1)}$ of the numerator $qS(a',b')$ of some Dedekind sum $S(a',b')$ with reduced denominator $q$.\n\nMore precisely, since $\\Lambda_q=\\Gamma_q$ but $\\lambda(a,t,t^{*})\\equiv 0 \\pmod{q}$ always, we have that $\\lambda(a,t,t^{*})\\pmod{q(q^2-1)}$ can take any value from $\\widetilde{\\Gamma}_q \\subseteq \\Z/q(q^2-1)\\Z$, defined by\n\\[\n\\widetilde{\\Gamma}_q = \\left\\{s \\in \\Z/q(q^2-1)\\Z: s \\equiv 0 \\pmod{q}, s \\pmod{q^2-1} \\in \\Gamma_q\\right\\},\n\\]\nwhere each element in $\\Gamma_q$ lifts uniquely to an element of $\\widetilde{\\Gamma}_q$ by the Chinese Remainder Theorem. Again because $\\Delta(\\ell)$ is periodic in $\\ell$ with period $q$, it can be considered as a map of $\\ell \\in (\\Z/q\\Z)^{\\times}$. Then if we define $\\phi: \\widetilde{\\Gamma}_q\\times (\\Z/q\\Z)^{\\times}\\to \\Z/(q(q^2-1))\\Z$ by\n\\[\n\\phi(\\lambda,\\ell) = \\lambda+\\Delta(\\ell)\\pmod{q(q^2-1)},\n\\]\nwe must have that the set of possible numerators $qS(a',b')$ is the image of $\\phi$.\n\nWe claim $\\phi$ is injective. Indeed, suppose there exist $\\lambda_1,\\lambda_2 \\in \\widetilde{\\Gamma}_q$ and $\\ell_1,\\ell_2 \\in (\\Z/q\\Z)^{\\times}$ such that \n\\[\n\\lambda_1+\\Delta(\\ell_1) \\equiv \\lambda_2 + \\Delta(\\ell_2)\\pmod{q(q^2-1)}.\n\\]\nSince $\\lambda_1\\equiv \\lambda_2 \\equiv 0 \\pmod{q}$, reducing $\\pmod{q}$ yields\n\\[\n\\Delta(\\ell_1)\\equiv \\Delta(\\ell_2) \\pmod{q}\n\\]\nso $\\ell_1\\equiv \\ell_2 \\pmod{q}$. This implies $\\Delta(\\ell_1) = \\Delta(\\ell_2)$, and so\n\\[\n\\lambda_1\\equiv \\lambda_2 \\pmod{q(q^2-1)}.\n\\]\nThen $\\phi$ is injective, so the number of residue classes of possible numerators $\\pmod{q(q^2-1)}$ is $|\\widetilde{\\Gamma}_q|\\varphi(q) = |\\Gamma_q|\\varphi(q)$, where $\\varphi$ denotes Euler's totient function. We claim that the number of residue classes $k$ satisfying the conditions of Theorem \\ref{qgood} is also $|\\Gamma_q|\\varphi(q)$. Indeed, note that $2\\nmid q$ if and only if $8 \\mid q^2-1$ and $3\\nmid  q$ if and only if $3\\mid q^2-1$. Since $\\gcd(q,q^2-1)=1$, by the Chinese Remainder Theorem the number of such residue classes $k\\pmod{q(q^2-1)}$ is\n\\[\n\\varphi(q)\\left(\\frac{q^2-1}{c_2(q)c_3(q)}\\right)\n\\]\nwhere\n\\[\nc_2(q) = \\begin{cases}\n1 & 2\\mid q\\\\\n4 & 2\\nmid q \\text{ and }q\\text{ nonsquare} \\\\\n8 & 2\\nmid q \\text{ and }q\\text{ square}\n\\end{cases}\n\\]\nand\n\\[\nc_3(q) = \\begin{cases}\n1 & 3 \\mid q\\\\\n3 & 3 \\nmid q.\n\\end{cases}\n\\]\nBut $8\\mid q^2-1$ if $2\\nmid q$ and $3\\mid q^2-1$ if $3\\nmid q$, while $2\\nmid q^2-1$ if $2\\mid q$ and $3\\nmid q^2-1$ if $3\\mid q$, so \n\\[\n|\\Gamma_q| = \\frac{q^2-1}{c_2(q)c_3(q)}\n\\]\nas well. Thus all possible residue classes $k \\pmod{q(q^2-1)}$ satisfying the conditions of Theorem $\\ref{qgood}$ are achievable as numerators of normalized Dedekind sums of denominator $q$. This, together with Theorem \\ref{finitemod}, establishes Conjecture \\ref{conj1}. \n\n\\end{proof}\n\\section{A Partial Resolution of Conjecture \\ref{conj2}}\n\\label{sec:partial}\n\nIn this section, we introduce identities that will allow us to prove Conjecture \\ref{conj2} (and therefore Conjecture \\ref{conj1}) in specific cases, such as $q$ even or $q$ an odd square divisible by $3$ or $5$. We also hypothesize an approach to proving Conjecture \\ref{conj2} in general and provide proof by computer verification for all $2\\le q \\le 200$.\n\n\\begin{defn}\nFor a given positive integer $q$ and an integer $a$ such that $\\gcd(q,a^2+1)=1$ let\n\\[\nf(a) = \\frac{(q^2-1) a}{a^2+1}-q S(aq,a^2+1).\n\\]\n\\end{defn}\nNote that by setting $t=a^2+1$, we have $f(a)\\pmod{q^2-1} \\in \\Lambda_q$ for any integer $a$ such that $\\gcd(q,a^2+1)=1$ (recall $f(a)$ is an integer). \n\nThe following identity reveals that $f(a)$ is a piecewise linear function depending on the residue class of $a\\pmod{q}$.\n\\begin{lemma}\n\\label{linear}\nIf an integer $a$ satisfies $\\gcd(q,a^2+1)=1$, then \n\\[\nf(a) = (g^2-1)a+S(a_1,q_1)+S(-a_1g^2-a_1^{*},q_1)\n\\]\nwhere $g = \\gcd(a,q)>0,a=ga_1,q=gq_1$, and $a_1a_1^{*}\\equiv 1 \\pmod{q_1}$.\n\\end{lemma}\n\\begin{proof}\nNote that $f(a)$ is an odd function, so without loss of generality we may assume that $a>0$. (The case $a=0$ is trivial.) By the reciprocity law, we have\n\\[\nS(aq,a^2+1)+S(a^2+1,aq) = \\frac{aq}{a^2+1}+\\frac{a^2+1}{aq}+\\frac{1}{aq(a^2+1)}-3.\n\\]\n\nNext, we apply the three-term relation (see \\cite{Girstmair98}) to $S(a^2+1,aq)$ and $S(a_1,q_1)$. Suppose integers $j$ and $k$ satisfy\n\\[\n-a_1j+q_1k=1,\n\\]\nand let\n\\[\nr = -aqk+(a^2+1)j.\n\\]\nThen \n\\[\nr = -ag(q_1k-a_1j)+j = -ag+j\n\\]\nand in particular\n\\[\nr\\equiv -a_1g^2-a_1^{*}\\pmod{q_1}.\n\\]\nBut \n\\[\n(a^2+1)q_1 - (aq)a_1=q_1\n\\]\nso\n\\[\nS(a^2+1,aq) = S(a_1,q_1)+S(-a_1g^2-a_1^{*},q_1) + \\frac{(aq)^2+2q_1^2}{aqq_1^2}-3.\n\\]\nCombining this with the reciprocity law directly implies the desired result.\n\\end{proof}\n\\begin{cor}\nIf $a$ is an integer such that $\\gcd(a^2+1,q)=1$ and $g = \\gcd(a,q)$, then \n\\[\nf(a+mq) = f(a)+mq(g^2-1)\n\\]\nfor any integer $m$.\n\\end{cor}\n\\begin{proof}\nNote that $\\gcd(a^2+1,q)=1$ implies $\\gcd((a+mq)^2+1,q)=1$, and furthermore $\\gcd(a,q) = \\gcd(a+mq,q)$. Thus the conclusion of Lemma \\ref{linear} holds for $a+mq$ with the same $g$ as $a$. But $S(a_1,q_1)+S(-a_1g^2-a_1^{*},q_1)$ only depends on the residue class of $a\\pmod{q}$, which implies the desired result.\n\\end{proof}\n\nAs a consequence, we have the following.\n\n\\begin{thm}\n\\label{mainresult}\nConjecture \\ref{conj2}, and thus Conjecture \\ref{conj1}, holds for $q$ even or a square divisible by $3$ or $5$.\n\\end{thm}\n\\begin{proof}\nFirst, suppose $q$ is even. If there exists an integer $a$ such that $\\gcd(a,q) = 2$ and $\\gcd(a^2+1,q)=1$, then $f(a+mq)=f(a)+3mq$ for any integer $m$. But $f(a+mq)\\pmod{q^2-1} \\in \\Lambda_q$ for any integer $m$, and $q$ is invertible $\\pmod{q^2-1}$, which implies $f(a)+3m \\pmod{q^2-1} \\in \\Lambda_q$ for all integers $m$. Now for $q$ even, $\\Gamma_q$ consists of multiples of $3\\pmod{q^2-1}$, so in particular $|\\Lambda_q|\\ge |\\Gamma_q|$. This, along with our prior assumption that $\\Lambda_q\\subseteq \\Gamma_q$, would imply $\\Lambda_q=\\Gamma_q$. \n\nSo it suffices in this case to show that there exists an integer $a$ such that $\\gcd(a,q)=2$ and $\\gcd(a^2+1,q)=1$. For each prime $p_i>2$ dividing $q$, we may choose a residue $r_i\\pmod{p_i}$ such that $r_i\\not\\equiv 0 \\pmod{p_i}$ and $r_i^2+1\\not\\equiv 0 \\pmod{p_i}$. This is because the two congruences only eliminate three of the $p_i$ residues for $p_i>3$, while if $p_i=3$, we may choose $r_i=1$. Then by the Chinese Remainder Theorem we may choose any $a$ such that\n\\[\na \\equiv 2 \\pmod{4}\n\\]\nand\n\\[\na \\equiv r_i\\pmod{p_i}\n\\]\nfor each $i$, which suffices.\n\nNow suppose $q$ is an odd square such that $3\\mid q$. In the same way, it suffices to find $a$ such that $\\gcd(a^2+1,q)=1$ and $\\gcd(a,q) = 3$, since $3^2-1=8$. Again for each $p_i>3$ dividing $q$, we may choose a remainder $r_i$ such that $r_i\\not\\equiv 0\\pmod{p_i}$ and $r_i^2+1\\not\\equiv 0\\pmod{p_i}$. By the Chinese Remainder Theorem, we can find our desired $a$ by taking $a \\equiv 3 \\pmod{9}$ and $a\\equiv r_i\\pmod{p_i}$.\n\nFinally, suppose $q$ is an odd square such that $5\\mid q$ and $3\\nmid q$. The argument is nearly identical to the case $3\\mid q$: it suffices to find $a$ such that $\\gcd(a^2+1,q)=1$ and $\\gcd(a,q)=5$, since $5^2-1=24$. Taking $a\\equiv 5 \\pmod{25}$ and applying the Chinese Remainder Theorem in the same way, we recover the desired value of $a$.\n\n\\end{proof}\n\nThe specification $t=a^2+1$ is too strong to prove Conjecture \\ref{conj2} in general. However, generalizing to $t=x^2+y^2$ yields enough flexibility to prove the conjecture for small values of $q$ (and possibly all $q$).\n\\begin{defn}\nFor a given integer $q\\ge 2$ and integers $x$ and $y$ such that $\\gcd(q,x^2+y^2)=\\gcd(x,y)=1$, let\n\\[\nh(x,y) = \\frac{(q^2-1)xy^{*}}{x^2+y^2}-qS(qxy^{*},x^2+y^2)\\pmod{q^2-1}\n\\]\nwhere $yy^{*}\\equiv 1 \\pmod{x^2+y^2}$. (Note that $y^{*}\\pmod{x^2+y^2}$ determines the expression $\\pmod{q^2-1}$, so the choice of $y^{*}$ doesn't matter.)\n\\end{defn}\nThis is exactly $\\lambda(a,t,t^{*})\\pmod{q^2-1}$ for $t = x^2+y^2$ and $a = xy^{*}$. Note that this generalizes $f(a)$, as $h(x,1)=f(x)\\pmod{q^2-1}$. In our case, this interpretation is useful for verifying small cases of Conjecture \\ref{conj2} computationally.\n\\begin{thm}\n\\label{computer}\nFor all integers $q$ such that $2\\le q\\le 200$, if $x$ and $y$ range over pairs of coprime integers such that $\\gcd(x^2+y^2,q)=1$, then $h(x,y)$ ranges over all of $\\Gamma_q$.\n\\end{thm}\n\\begin{proof}\nA naive computer search over pairs of coprime positive integers $x$ and $y$ suffices for the values of $q$ not already covered by Theorem \\ref{mainresult}. Our Sage program terminates in under an hour with $x+y$ achieving the largest value when $(x,y) = (1576,1511)$ for $q=189$.\n\\end{proof}\n\\begin{cor}\nConjectures \\ref{conj2} and \\ref{conj1} hold for all positive integers $2 \\le q\\le 200$.\n\\end{cor}\n\nFor future work, we would hope to see a proof for Conjecture \\ref{conj2}, and therefore Conjecture \\ref{conj1}, in full generality. We conjecture that setting $t=x^2+y^2$ and $a=xy^{*}$ as in Theorem \\ref{computer} is enough to do so:\n\\begin{conj}\nFix $q\\ge 2$. If $x$ and $y$ vary over pairs of coprime integers such that $\\gcd(q,x^2+y^2)=1$, then $h(x,y)$ ranges over $\\Gamma_q$.\n\\end{conj}\nOnce again, as $h(x,y) \\in \\Lambda_q$, this would imply Conjecture \\ref{conj2}.\n\nIn fact, we believe that $h(x,y)$ can achieve all values in $\\Gamma_q$ with one fewer parameter. By setting $t = (x^2+1)(q^2+1) = (x+q)^2+(xq-1)^2$, we have\n\\begin{conj}\nFix $q\\ge 2$. If $x$ ranges over all integers such that $\\gcd(x+q,xq-1)=1$ and $\\gcd(x^2+1,q)=1$, then $h(x+q,xq-1)$ ranges over $\\Gamma_q$.\n\\end{conj}\n\nMore generally, we conjecture that for any positive integer $r$, we can set $t = (x^2+1)(r^2q^2+1) = (x+rq)^2+(xrq-1)^2$ and get the same result.\n\\begin{conj}\nFix $q\\ge 2$ and $r\\ge 1$. If $x$ ranges over all integers such that $\\gcd(x+rq,xrq-1)=1$ and $\\gcd(x^2+1,q)=1$, then $h(x+rq,xrq-1)$ ranges over $\\Gamma_q$.\n\\end{conj}\n\n\\section{Acknowledgments}\nThis work was supported by NSF grant DMS-1659047 as part of the Duluth Research Experience for Undergraduates (REU). The author would like to thank Joe Gallian for supervising the program, suggesting the problem, and providing helpful commentary on drafts of this paper.\n\n\\bibliographystyle{plain}\n", "meta": {"timestamp": "2018-12-27T02:31:46", "yymm": "1801", "arxiv_id": "1801.00517", "language": "en", "url": "https://arxiv.org/abs/1801.00517", "abstract": "Let $S(a,b)$ denote the normalized Dedekind sum. We study the range of possible values for $S(a,b)=\\frac{k}{q}$ with $\\gcd(k,q)=1$. Girstmair proved local restrictions on $k$ depending on $q\\pmod{12}$ and whether $q$ is a square and conjectured that these are the only restrictions possible. We verify the conjecture in the cases $q$ even, $q$ a square divisible by $3$ or $5$, and $2\\le q\\le 200$ (the latter by computer), and provide progress towards a general approach.", "subjects": "Number Theory (math.NT)", "title": "Dedekind Sums with Even Denominators", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.984336353126336, "lm_q2_score": 0.8596637451167997, "lm_q1q2_score": 0.8461982757831986}}
{"url": "https://arxiv.org/abs/2112.02181", "title": "Projections onto hyperbolas or bilinear constraint sets in Hilbert spaces", "abstract": "Sets of bilinear constraints are important in various machine learning models. Mathematically, they are hyperbolas in a product space. In this paper, we give a complete formula for projections onto sets of bilinear constraints or hyperbolas in a general Hilbert space.", "text": "\\section{Introduction}\nThroughout this paper, we assume that\n\\begin{empheq}[box=\\mybluebox]{equation}\n    \\text{$X$ is a Hilbert space\n   \n    with inner product\n    $\\scal{\\cdot}{\\cdot}\\colon X\\times X\\to\\ensuremath{\\mathbb R}$, }\n\\end{empheq}\nand induced norm $\\|\\cdot\\|$. In various learning models \\cite{elser2019learning,elser2017products}, Elser utilizes\nprojections onto the \\emph{bilinear constraint} set:\n\\begin{empheq}[box=\\mybluebox]{equation}\n\\label{e:defC}\nC_\\gamma := \\menge{(x,y)\\in X\\times\nX}{\\scal{x}{y}=\\gamma}\n\\end{empheq}\nwhere $\\gamma\\in\\ensuremath{\\mathbb R}$ is a fixed constant. Mathematically speaking, up to a rotation, this is just a hyperbola or\nquadratic surface in $X\\times X$. Finding projections onto quadratic curves or surfaces algorithmically\nhave many practical applications;\nsee, e.g., \\cite{quadrics, conics, elser2019learning, elser2017products}.\nIf $\\gamma=0$ in $C_{\\gamma}$, then $C_0$ becomes the set `cross' in $X\\times X$.\nThe projection formula for the `cross' has been thoroughly investigated \\cite{cross}.\nIn \\cite{elser2019learning,elser2017products}, although Elser has provided some\nresults on projections\nonto $C_{\\gamma}$ only when $X=\\ensuremath{\\mathbb R}^n$, complete mathematical details are not presented.\n\n\\emph{The goal of this paper is to give a complete analysis of the projection onto the set $C_{\\gamma}$ with $\\gamma\\neq 0$,\nand extend the results to a general Hilbert space.}\n\n\n\n\n\\noindent The remainder of the paper is organized as follows. \\cref{sec:aux} gives\nsome general properties of bilinear constraint sets. In \\cref{section:hyperbola} we\nfocus on full mathematical details\nfor the existence and explicit formula of projections onto hyperbolas. Using results\nfrom \\cref{section:hyperbola}, in\nSections~\\ref{sec:positive:gam}--\\ref{sec:negative:gam} we provide explicit formulas for\nprojections onto sets of bilinear constraints.\nOur notation is standard and follows largely \\cite{BC2017,Rock98}.\n\\section{General properties of $C_{\\gamma}$}\n\\label{sec:aux}\nIn an infinite-dimensional space, the existence of projection onto a set\noften requires the weak closedness of the set.\nOur first result says that\nalthough the set $C_{\\gamma}$ is norm closed it is not weakly closed in $X\\times X$.\n\\begin{proposition}\\label{p:weak}\n\nThe set $C_{\\gamma}$ is closed in the\nnorm\ntopology but\nnot closed in the weak topology in $X\\times X$.\nIn fact, $\\overline{C_{\\gamma}}^{\\rm weak}=X\\times X$.\n\\end{proposition}\n\n\\begin{proof} Evidently, $C_{\\gamma}$ is norm closed. We show that\n$\\overline{C_{\\gamma}}^{\\rm weak}=X\\times X$.\nLet $(x,y)\\in X\\times X$. Consider $S=\\mbox{span}\\{x,y\\}$.\nThe orthogonal decomposition theorem gives $X= S\\oplus S^{\\perp}$.\nBecause $X$ is infinite-dimensional and $S$ is at most two-dimensional, $S^{\\perp}$ is infinite dimensional, so\nany orthonormal base of $S^{\\perp}$ must have a sequence $(e_{n})_{\\ensuremath{{n\\in{\\mathbb N}}}}$ which converges weakly to $0$, i.e.,\n$e_{n}\\rightharpoonup 0$; see, e.g., \\cite{kreyszig}.\nAlso $e_{n}\\perp x$ and\n$e_{n}\\perp y.$\nSet $\\xi =\\gamma-\\langle x, y\\rangle$. Then\n\\begin{align}\n\\scal{x-\\xi e_{n}}{y-e_{n}} &=\n\\scal{x}{y}-\\scal{x}{e_{n}}-\\xi\\scal{e_{n}}{y}+\\xi\\scal{e_{n}}{e_{n}}\\\\\n&=\\scal{x}{y}+\\xi  =\\gamma,\n\\end{align}\nso $(x-\\xi e_{n}, y-e_{n})\\in C_{\\gamma}$.\nSince $(x-\\xi e_{n}, y-e_{n}) \\rightharpoonup (x,y)$, we have\n $(x, y)\\in \\overline{C_{\\gamma}}^{\\rm weak}$.\nBecause $(x,y)\\in X\\times X$ was arbitrary, we conclude that\n$X\\times X\\subseteq \\overline{C}^{\\rm weak}$.\n\\end{proof}\n\nProposition~\\ref{p:weak} indicates that finding $P_{C_{\\gamma}}$, i.e., projections onto $C_{\\gamma}$,\nmight be complicated in a general Hilbert space.\nThis seemingly difficult issue can be completely avoided by utilizing the structures of\nthe optimization problem. See Subsection~\\ref{s:existence} below.\n\n\n\\begin{comment}\nThe following facts work in Hilbert space.\n\n\\begin{fact}\n\\label{f:Bert1}\nIf $x^*$ is a local minimizer of $f(x)$ subject to $h_1(x)=0,\\ldots,\nh_m(x)=0$ and the gradients $\\nabla h_1(x^*),\\ldots,\\nabla h_m(x^*)$\nare linearly independent, then\nthere a unique vector $(\\lambda_1^*,\\ldots,\\lambda_m^*)$ such that\n\\begin{equation}\n\\nabla f(x^*) + \\sum_{i=1}^{m}\\lambda_i \\nabla h_i(x^*) = 0.\n\\end{equation}\n\\end{fact}\n\\begin{proof}\nSee \\cite[Proposition~4.1.1]{Bertsekas}.\n\\end{proof}\n\n\\begin{fact}\n\\label{f:Bert2}\nGiven $f$ and $h$ are twice continuously differentiable, and let $x^{*} \\in \\ensuremath{\\mathbb R}^n$ and $\\lambda^{*} \\in \\ensuremath{\\mathbb R}^m$ satisfy\n\\begin{align}\n   \\nabla_x L(x^{*},\\lambda^{*}) = 0, ~ \\nabla_{\\lambda} L(x^{*},\\lambda^{*}) = 0 \\\\\n   y'\\nabla^{2}_{xx}L(x^{*},\\lambda^{*})y>0, ~~ \\forall y\\neq 0 ~\\text{with}~\\nabla h(x^{*})'y = 0.\n\\end{align}\nThen $x^*$ is a strict local minimum of $f(x)$ subject to $h(x)=0$. In fact, there exist scalers $\\delta>0$ and $\\epsilon>0$ such that\n\\begin{align}\n f(x) \\geq f(x^{*})+\\frac{\\delta}{2}\\norm{x-x^{*}}^2, \\quad \\forall x ~\\text{with}~ h(x)= 0~\\text{and}~\\norm{x-x^{*}} < \\epsilon\n\\end{align}\n\\end{fact}\n\\begin{proof}\nSee \\cite[Proposition~4.2.1]{Bertsekas}.\n\\end{proof}\\\\\nNote that the above sufficient conditions do not include regularity of the vector $x^{*}$.\n\\end{comment}\n\n\n\n\nOur next result says that when $\\gamma\\neq 0$, locally around\nthe set $C_{\\gamma}$ the projection onto the set is always single-valued.\n\\begin{proposition}\\label{p:regular}\n Let $\\gamma\\neq 0$, and $C_{\\gamma}\n= \\menge{(x,y)\\in X\\times X}{\\scal{x}{y}=\\gamma}.$\nFor every $(x,y)\\in C_{\\gamma}$,\nthe the following hold:\n\\begin{enumerate}\n\\item\\label{i:prox1}\n$C_{\\gamma}$ is prox-regular at $(x,y)$.\n\\item\\label{i:prox2}\nThere exists a neighborhood of $(x,y)$\non which\nthe projection mapping onto $C_{\\gamma}$ is monotone and Lipschitz continuous.\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof}\n\\ref{i:prox1}: Write\n$\\iota_{C_{\\gamma}}(x,y)=\\iota_{\\{\\gamma\\}}(h(x,y))$ where $h(x,y)=\\scal{x}{y}$. Let\n$(x,y)\\in C_{\\gamma}$. Then\n$\\nabla h(x,y)=(y,x) \\neq (0,0)$ because of $\\gamma\\neq 0$.\n Being a composition of a convex function $\\iota_{\\{\\gamma\\}}$\nand a twice\ndifferentiable function $h$ that is qualified at $(x,y)$, \\cite[Proposition 2.4]{thibault} shows that\n$\\iota_{C_{\\gamma}}$ is prox-regular at $(x,y)$, so is $C_{\\gamma}$.\n\n\\ref{i:prox2}: Apply \\ref{i:prox1} and \\cite[Proposition 4.4]{thibault}.\n\\end{proof}\n\nObserve that  Proposition~\\ref{p:regular}\nalso follow from \\cite[Proposition 13.32]{Rock98} and \\cite[Exercise 13.38]{Rock98} when $X=\\ensuremath{\\mathbb R}^n$; and that\nwhen $\\gamma=0$, $C_{0}$ is not prox-regular at $(0,0)$. Although the projection\nexists locally around $C_{\\gamma}$, it is still not clear for the global existence.\n\n\n\n\\section{Projections onto hyperbolas}\n\\label{section:hyperbola}\nFor ease of analysis, we start with\n\\begin{empheq}{equation*}\nC_{\\gamma} := \\menge{(x,y)\\in X\\times X}{h(x,y):=\\scal{x}{y}-\\gamma=0}\n\\end{empheq}\nwhere $\\gamma>0$.\nOur goal is to find the projection formula\n$P_{C_{\\gamma}}(x_0,y_{0})$ for every $(x_{0},y_{0})\\in X\\times X$. That is,\n\\begin{empheq}{equation*}\n\\text{minimize} \\quad f(x,y):=\\|x-x_{0}\\|^2+\\|y-y_{0}\\|^2\n\\quad\\text{subject to} \\quad (x,y) \\in C_{\\gamma}.\\tag{P}\n\\end{empheq}\n\n\\subsection{Auxiliary problems and existence of projections}\n\\label{s:existence}\n\\noindent To determine the projection operator $\\proj{C_{\\gamma}}$ of the set $C_{\\gamma}$,\nwe shall introduce two\nequivalently\nreformulated problems.\nFirst, for every $(u_{0},v_{0})\\in X\\times X$, we solve the problem\n\\begin{empheq}{equation*}\n\\text{minimize} \\quad f_{1}(u,v):=\\|u-u_{0}\\|^2+\\|v-v_{0}\\|^2\n\\quad\\text{subject to} \\quad (u,v) \\in \\ensuremath{\\tilde{C}}_1, \\tag{$\\tilde{P}_{1}$}\n\\end{empheq}\nwhere  $\\ensuremath{\\tilde{C}}_1:= \\menge{(u,v)\\in \\ensuremath{{X}} \\times \\ensuremath{{X}}}{h_{1}(u,v):=\\norm{u}^2 -\\norm{v}^2-2=0}.$\nNext, for every $(\\ensuremath{\\tilde{u}}_{0},\\ensuremath{\\tilde{v}}_{0})\\in X\\times X$ we solve the problem\n\\begin{empheq}{equation*}\n\\text{minimize} \\quad f_{\\gamma}(\\tilde{u},\\tilde{v}):=\\|\\tilde{u}-\\tilde{u}_{0}\\|^2+\\|\\tilde{v}-\n\\tilde{v}_{0}\\|^2\n\\quad\\text{subject to} \\quad (\\tilde{u},\\tilde{v}) \\in \\ensuremath{\\tilde{C}}_{\\gamma}, \\tag{$\\tilde{P}_{\\gamma}$}\n\\end{empheq}\nwhere $\\ensuremath{\\tilde{C}}_{\\gamma}:= \\menge{(\\tilde{u},\\tilde{v})\\in \\ensuremath{{X}} \\times \\ensuremath{{X}}}{h_{\\gamma}(\n\\tilde{u},\\tilde{v}):=\\norm{\\tilde{u}}^2 -\\norm{\\tilde{v}}^2-2\\gamma=0}$.\nBoth $\\ensuremath{\\tilde{C}}_{1}$ and $\\ensuremath{\\tilde{C}}_{\\gamma}$ are hyperbolas.\n$\\tilde{P}_{1}$ and $\\tilde{P}_{\\gamma}$ solve for projections $P_{\\ensuremath{\\tilde{C}}_{1}}$ and $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}$ respectively,\nand their connections to\n$P_{C_{\\gamma}}$ are given by the \\cref{p:reformulate} below. Recall\n\\begin{definition}[Rotation with an angle $\\phi$]\n\\label{f:rotation}\nA change of coordinates $(u,v) \\in \\ensuremath{{X}} \\times \\ensuremath{{X}}$, by rotation through an angle $\\phi$, is defined by\n$$\n\\begin{bmatrix}x\\\\y\\end{bmatrix}=\\begin{bmatrix}\\cos\\phi\\ensuremath{\\operatorname{Id}} & -\\sin\\phi\\ensuremath{\\operatorname{Id}} \\\\\n\\addlinespace\n\\sin\\phi\\ensuremath{\\operatorname{Id}} & \\cos\\phi\\ensuremath{\\operatorname{Id}} \\end{bmatrix}\n\\begin{bmatrix}u\\\\v\\end{bmatrix}.$$\n\\end{definition}\nPut\n$$A_{\\phi}:=\\begin{bmatrix}\\cos\\phi\\ensuremath{\\operatorname{Id}} & -\\sin\\phi\\ensuremath{\\operatorname{Id}} \\\\\n\\addlinespace\n\\sin\\phi\\ensuremath{\\operatorname{Id}} & \\cos\\phi\\ensuremath{\\operatorname{Id}} \\end{bmatrix}.$$\nThen $A_{\\phi}A_{-\\phi}=\\ensuremath{\\operatorname{Id}}=A_{-\\phi}A_{\\phi}$ and $A^{-1}_{\\phi}=A_{-\\phi}$.\nThe relationships among $P_{C_{\\gamma}}, P_{\\ensuremath{\\tilde{C}}_{\\gamma}}$ and $P_{\\ensuremath{\\tilde{C}}_{1}}$ are summarized below.\n\\begin{proposition}\\label{p:reformulate}\nThe following hold:\n\\begin{enumerate}\n\\item\\label{i:scaling} $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}=\\sqrt{\\gamma}P_{\\ensuremath{\\tilde{C}}_{1}}(\\ensuremath{\\operatorname{Id}}/\\sqrt{\\gamma}).$\n\\item\\label{i:rotate} $P_{C_{\\gamma}}=A_{\\pi/4}P_{\\ensuremath{\\tilde{C}}_{\\gamma}}A_{-\\pi/4}.$\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof}\n\\ref{i:scaling}: $(\\tilde{P}_{1})$ is equivalent to $(\\tilde{P}_{\\gamma})$\nby a change of variables of scaling\n$\\ensuremath{\\operatorname{Id}}/\\sqrt{\\gamma}$. Indeed, using\n$$\n\\begin{bmatrix}\nu\\\\\nv\n\\end{bmatrix}\n=\n\\frac{1}{\\sqrt{\\gamma}}\n\\begin{bmatrix}\n\\tilde{u}\\\\\n\\tilde{v}\n\\end{bmatrix}, \\text{ and } \\begin{bmatrix}\nu_{0}\\\\\nv_{0}\n\\end{bmatrix}\n=\n\\frac{1}{\\sqrt{\\gamma}}\n\\begin{bmatrix}\n\\ensuremath{\\tilde{u}}_{0}\\\\\n\\ensuremath{\\tilde{v}}_{0}\n\\end{bmatrix}\n$$\n$h_{\\gamma}(\\tilde{u},\\tilde{v})=0$ becomes $h_{1}(u,v)=0$, and\n$f_{\\gamma}(\\tilde{u},\\tilde{v})$ becomes\n$\\gamma f_{1}(u,v)=\\gamma(\\|u-u_{0}\\|^2+\\|v-v_0\\|^2)$.\n\n\n\\ref{i:rotate}: $(P)$ is equivalent to $(\\tilde{P}_{\\gamma})$ by a change of variables of\nrational $A_{\\pi/4}$. Indeed, with\n$$\\begin{bmatrix}\nx\\\\\ny\n\\end{bmatrix}\n  =A_{\\pi/4}\\begin{bmatrix}\n  \\tilde{u}\\\\\n  \\tilde{v}\n  \\end{bmatrix}, \\text{ and }\n  \\begin{bmatrix}\nx_{0}\\\\\ny_{0}\n\\end{bmatrix}\n  =A_{\\pi/4}\\begin{bmatrix}\n  \\tilde{u}_{0}\\\\\n  \\tilde{v}_{0}\n  \\end{bmatrix}\n  $$\n the objective $f(x,y) = \\|x-x_0\\|^2 + \\|y-y_0\\|^2=\\|(x,y)-(x_{0},y_0)\\|^2$ can be rewritten as\n \\begin{align}\n   f(x,y) &=\n \n\\norm{A_{\\frac{\\pi}{4}}\\begin{bmatrix}\n           \\tilde{u}-\\tilde{u}_0 \\\\\n          \\tilde{v}-\\tilde{v}_0 \\end{bmatrix}\n     }^2 \\nonumber =\n     \\begin{bmatrix}\n           \\tilde{u}-\\tilde{u}_{0} \\\\\n          \\tilde{v}-\\tilde{v}_{0} \\end{bmatrix}^{T}A_{\\frac{\\pi}{4}}^{T}A_{\\frac{\\pi}{4}}\\begin{bmatrix}\n           \\tilde{u}-\\tilde{u}_{0} \\\\\n          \\tilde{v}-\\tilde{v}_{0} \\end{bmatrix} \\\\\n&\n=\\norm{\\tilde{u}-\\tilde{u}_0}^2 + \\norm{\\tilde{v}-\\tilde{v}_0}^2, \\nonumber\\label{new_obj}\n  \\end{align}\n  and\n  $h(x,y)=\\scal{x}{y}-\\gamma=0$ becomes $h_{\\gamma}(\\tilde{u},\\tilde{v})=\n  \\|\\tilde{u}\\|^2-\\|\\tilde{v}\\|^2-2\\gamma=0$.\n\\end{proof}\n\nIn view of Proposition~\\ref{p:reformulate}\\ref{i:rotate}, to see\nthat $P_{C_{\\gamma}}(x,y)\\neq \\varnothing$ for every $(x,y)\\in X\\times X$,\nthe following observation is crucial.\n\n\\begin{proposition}\\label{p:two:dim}\nFor every $(\\ensuremath{\\tilde{u}}_0,\\ensuremath{\\tilde{v}}_0) \\in \\ensuremath{{X}}\\times\\ensuremath{{X}}$,\nthe minimization problem\n\\begin{eqnarray}\n\\text{minimize} & \\quad f_{\\gamma}(\\ensuremath{\\tilde{u}},\\ensuremath{\\tilde{v}}) =\\|\\tilde{u}-\\tilde{u}_{0}\\|^2+\\|\\tilde{v}-\n\\tilde{v}_{0}\\|^2 \\label{e:simple1}\\\\\n\\text{subject to} & \\quad  h_{\\gamma}(\\ensuremath{\\tilde{u}},\\ensuremath{\\tilde{v}}) =\\norm{\\tilde{u}}^2 -\\norm{\\tilde{v}}^2-2\\gamma =0\n\\label{e:simple2}\n\\end{eqnarray}\nalways has a solution, i.e., $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\ensuremath{\\tilde{u}}_{0},\\ensuremath{\\tilde{v}}_{0})\\neq\\varnothing$.\n\\end{proposition}\n\n\\begin{proof} We shall illustrate only the case $\\ensuremath{\\tilde{u}}_{0}\\neq 0, \\ensuremath{\\tilde{v}}_{0}\\neq 0$, since\nthe other cases are similar. We claim that the optimization problem is essentially\n$2$-dimensional. To this end, we expand\n\\begin{align}\nf_{\\gamma}(\\ensuremath{\\tilde{u}},\\ensuremath{\\tilde{v}}) &\n=\\underbrace{\\|\\ensuremath{\\tilde{u}}\\|^2-2\\scal{\\ensuremath{\\tilde{u}}}{\\ensuremath{\\tilde{u}}_{0}}+\\|\\ensuremath{\\tilde{u}}_{0}\\|^2}+\\underbrace{\\|\\ensuremath{\\tilde{v}}\\|^2-2\\scal{\\ensuremath{\\tilde{v}}}{\\ensuremath{\\tilde{v}}_{0}}+\\|\\ensuremath{\\tilde{v}}_{0}\\|^2}.\n\\end{align}\nThe constraint\n$$h_{\\gamma}(\\ensuremath{\\tilde{u}},\\ensuremath{\\tilde{v}}):=\\norm{\\tilde{u}}^2 -\\norm{\\tilde{v}}^2-2\\gamma =0$$\nmeans that only the norms $\\|\\ensuremath{\\tilde{u}}\\|$ and $\\|\\ensuremath{\\tilde{v}}\\|$ matter.\nWith $\\|\\ensuremath{\\tilde{u}}\\|$ fixed, the Cauchy-Schwarz inequality in a Hilbert space, see, e.g., \\cite{kreyszig},\nshows that $\\ensuremath{\\tilde{u}}\\mapsto\\scal{\\ensuremath{\\tilde{u}}}{\\ensuremath{\\tilde{u}}_{0}}$\nwill be larger so that\nthe first underlined part in $f_{\\gamma}$ will be smaller when\n$\\ensuremath{\\tilde{u}}$ and $\\ensuremath{\\tilde{u}}_{0}$ are positively co-linear, i.e, $\\ensuremath{\\tilde{u}}=\\alpha\\ensuremath{\\tilde{u}}_{0}$ for some\n$\\alpha\\geq 0$. Similarly, for fixed $\\|v\\|$ the second underlined part in\n$f_{\\gamma}$ will be smaller when\n$\\ensuremath{\\tilde{v}}=\\beta\\ensuremath{\\tilde{v}}_{0}$ for some $\\beta\\geq 0$. It follows that the optimization\nproblem given by \\eqref{e:simple1}-\\eqref{e:simple2} is equivalent to\n\\begin{align}\n\\text{minimize} &\\quad g(\\alpha,\\beta):=(1-\\alpha)^2\\|\\tilde{u}_{0}\\|^2+(1-\\beta)^2\\|\n\\tilde{v}_{0}\\|^2\\label{e:sunday1}\\\\\n\\text{subject to} & \\quad g_{1}(\\alpha,\\beta):=\\alpha^2\\norm{\\tilde{u}_{0}}^2 -\n\\beta^2\\norm{\\tilde{v}_{0}}^2-2\\gamma =0, \\alpha\\geq 0, \\beta\\geq 0.\\label{e:sunday2}\n\\end{align}\nBecause $g:\\ensuremath{\\mathbb R}^2\\rightarrow\\ensuremath{\\mathbb R}$ is continuous and coercive, and $g_{1}:\\ensuremath{\\mathbb R}^2\\rightarrow\\ensuremath{\\mathbb R}$ is\ncontinuous, we conclude that the optimization problem given by\n\\eqref{e:sunday1}-\\eqref{e:sunday2} has a solution.\n\\end{proof}\n\n\nWe are now ready for the investigation of projections onto $\\ensuremath{\\tilde{C}}_{1}, \\ensuremath{\\tilde{C}}_{\\gamma}$ and $C_{\\gamma}$.\n\n\n\\subsection{Finding the projection $P_{\\ensuremath{\\tilde{C}}_{1}}$}\nNote that\n\\begin{equation}\n\\label{e:ngradients}\n\\nabla f_{1}(u,v) = (2(u-u_0),2(v-v_0))\n\\quad\\text{and}\\quad\n\\nabla h_{1}(u,v) = (2u,-2v).\n\\end{equation}\nBy \\cite[Proposition~4.1.1]{Bertsekas}, every solution of ($\\tilde{P}_{1}$) satisfies\nthe necessarily conditions, i.e., the KKT system of ($\\tilde{P}_{1}$), given by\n\\begin{align}\n(1+\\lambda) u = u_0 \\label{eq:l1}\\\\\n(1 - \\lambda) v = v_0 \\label{eq:l2}\\\\ \\|u\\|^2-\\|v\\|^2-2=0\\label{eq:l3}.\n\\end{align}\n\\begin{lemma}\\label{l:bounds}\nLet $u_0\\neq 0,v_0\\neq 0$. If $(u, v, \\lambda)$ verifies \\eqref{eq:l1}--\\eqref{eq:l3} and\n$(u,v)$ is an optimal solution of problem ($\\tilde{P}_{1}$), then\nthe Lagrange multiplier $\\lambda$\n satisfies $|\\lambda|<1$.\n\\end{lemma}\n\\begin{proof}\nThe constraint set\n$$\\|u\\|^2-\\|v\\|^2=2,$$\nhas a special structure: replacing $u$ by $-u$, or $v$ by $-v$\nthe constraint is still verified.\nAlso consider\n$$f(u,v)=\\|u-u_{0}\\|^2+\\|v-v_{0}\\|^2=\\|u_{0}\\|^2-2\\scal{u_{0}}{u}+\n\\|u\\|^2+\\|v_0\\|^2-2\\scal{v_{0}}{v}+\\|v\\|^2,$$\nGiven $u_{0} \\text{ and }\nv_{0}$, for fixed $\\|u\\|$ and $\\|v\\|$, $f(u,v)$ becomes smaller if one choose\n$\\scal{u_{0}}{u}\\geq 0$, and\n$\\scal{v_{0}}{v}\\geq 0$. Indeed, one can do so by replacing $u$ by $-u$ or $v$ by $-v$ if needed.\nNow by \\eqref{eq:l1} and \\eqref{eq:l2},\n$$(1+\\lambda)\\scal{u_{0}}{u}=\\|u_{0}\\|^2, \\text{ and }\n(1-\\lambda)\\scal{v_{0}}{v}=\\|v_{0}\\|^2.$$\nBecause $u_{0}\\neq 0, v_{0}\\neq 0$, we have $\\scal{u_{0}}{u}>0, \\scal{v_{0}}{v}>0$ so that\n$1+\\lambda>0, 1-\\lambda>0$. Hence\n$|\\lambda|<1.$\n\\end{proof}\n\n\\begin{proposition}\n\\label{p:2011d4.1}\nLet $u_0\\neq0, v_0\\neq 0$. Define $p \\coloneqq\\norm{u_0}^2 - \\norm{v_0}^2 $ and $q \\coloneqq\\norm{u_0}^2 + \\norm{v_0}^2$.\nSuppose that $u,v\\in\\ensuremath{{X}}$ and $\\lambda\\in\\ensuremath{\\mathbb R}$ verify \\eqref{eq:l1}-\\eqref{eq:l3} and $(u,v)$ is an optimal solution\nto ($\\tilde{P}_{1}$).\nThen\nthe following hold:\n\\begin{enumerate}\n\\item\n\\label{p:2011411}\n$u = \\frac{u_0}{(1+\\lambda)}$, $v = \\frac{v_0}{(1-\\lambda)}$, and\n$(1-\\lambda)^2\\norm{u_0}^2 - (1+\\lambda)^2\\norm{v_0}^2\n= 2(1-\\lambda^2)^2$.\n\n\\item\n\\label{p:2011412}\nThe objective function has\n$\nf_{1}(u,v) = \\lambda^2\\bigg(\\frac{\\|u_0\\|^2}{(1+\\lambda)^2}+\\frac{\\|v_0\\|^2}{(1-\\lambda)^2} \\bigg).\n$\n\\item\n\\label{p:2011413}\n$\\lambda$ is the unique solution of\n\\begin{equation}\n\\label{eq:main}\nH(\\lambda) \\coloneqq \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} -1  = 0\n\\end{equation}\nin $\\left]-1,1\\right[$.\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof}\n\\cref{p:2011411}: Because $u_0\\neq0, v_0\\neq0$, we obtain $|\\lambda|<1$,\n$u = \\frac{u_0}{(1+\\lambda)}$ and $v = \\frac{v_0}{(1-\\lambda)}$ by\nLemma~\\ref{l:bounds} and \\eqref{eq:l1}-\\eqref{eq:l2}.\nThen we have the following equivalences\n\\begin{align}\n\\norm{u}^2 - \\norm{v}^2 =2\n&\\Leftrightarrow\n\\frac{\\norm{u_0}^2}{(1+\\lambda)^2} - \\frac{\\norm{v_0}^2}{(1-\\lambda)^2} = 2 \\\\\n&\\Leftrightarrow\n(1-\\lambda)^2\\norm{u_0}^2  - (1+\\lambda)^2 \\norm{v_0}^2 =2(1-\\lambda)^2(1+\\lambda)^2\\\\\n&\\Leftrightarrow\n(1-\\lambda)^2\\norm{u_0}^2  - (1+\\lambda)^2 \\norm{v_0}^2 =2(1-\\lambda^2)^2.\n\\end{align}\n\\cref{p:2011412}: Substitute $u = \\frac{u_0}{(1+\\lambda)}$ and $v = \\frac{v_0}{(1-\\lambda)}$ in $f_{1}$.\\\\\n\\cref{p:2011413}:\nBy \\cref{p:2011411}, we have\n\\begin{equation}\n\\label{e:2011416}\n(1-\\lambda)^2\\norm{u_0}^2 - (1+\\lambda)^2 \\norm{v_0}^2 =2(1-\\lambda^2)^2.\n\\end{equation}\nSince\n\\begin{align}\n&\\hspace{-1cm}\n(1-\\lambda)^2\\norm{u_0}^2 - (1+\\lambda)^2 \\norm{v_0}^2\\\\\n&=\n\\big(1+\\lambda^2-2\\lambda\\big)\\|u_0\\|^2 - \\big(1+\\lambda^2+2\\lambda\\big)\\|v_0\\|^2\\\\\n&=\n\\lambda^2\\big(\\|u_0\\|^2-\\|v_0\\|^2\\big)+\\big(\\|u_0\\|^2-\\|v_0\\|^2\\big)-2\\lambda\\big(\\|u_0\\|^2+\\|v_0\\|^2\\big) \\\\\n&=\\lambda^2 p+p-2\\lambda q, \\label{e:num}\n\\end{align}\nusing \\cref{e:num} on left side of \\cref{e:2011416}, we obtain $\\lambda^2 p+p-2\\lambda q =2(1-\\lambda^2)^2$,\na univariate quartic equation in $\\lam$, equivalently,\n\\begin{align}\n\\label{eq:quartic1}\n H(\\lambda) \\coloneqq \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} -1  = 0.\n\\end{align}\nWe show that \\cref{eq:quartic1} has a unique solution in $]-1,1[$.\nBecause $u_0\\neq0, v_0\\neq0$,\nwe know that $q >|p|$,\nand using it along-with $|\\lambda|<1$ from \\cref{l:bounds},\nwe get $p\\lambda \\leq |p||\\lambda|\\leq q|\\lambda|$, and\n\\begin{align}\nH'(\\lambda)\n&= \\frac{1}{(1-\\lambda^2)^3}\\big(-q(1+3\\lambda^2) + p(\\lambda^3+3\\lambda)\\big)\\\\\n&\\leq\\frac{1}{(1-\\lambda^2)^3}\\big(-q(1+3\\lambda^2) + |p||\\lambda|(\\lambda^2+3)\\big)\\\\\n&\\leq\\frac{1}{(1-\\lambda^2)^3}\\big(-q(1+3\\lambda^2)+ q|\\lambda|(\\lambda^2+3)\\big)\\\\\n&= \\frac{q}{(1-\\lambda^2)^3}\\big((|\\lambda|-1)^3\\big)\n\\\\\n&= \\frac{-q}{((1-|\\lambda|)(1+|\\lambda|))^3}\\big((1-|\\lambda|)^3\\big)\n\\\\ &= \\frac{-q}{(1+|\\lambda|)^3}.\n\\end{align}\nSince $q = \\|u_0\\|^2 + \\|v_0\\|^2 > 0$ and $(1+|\\lambda|)^3 > 0$, we have $H'(\\lambda)<0$ on $]-1,1[$, which implies  $H(\\lambda)$\nis strictly decreasing on $]-1,1[$. Notice that $\\lambda = \\pm 1$ are vertical asymptotes because\nat $\\lambda = \\pm 1$ the numerator of $H(\\lambda)$ is, $H(-1) = 2(p+q)=4\\|u_0\\|^2 > 0$ and $H(1) = 2(p - q) = - 4 \\|v_0\\|^2 < 0$, so\nthat $\\lim_{\\lambda \\to 1^{-}} H(\\lambda) = - \\infty$ and $\\lim_{\\lambda \\to -1^{+}} H(\\lambda) = + \\infty$. Since\n$H$ is continuous, strictly decreasing, and its range is $]-\\infty, \\infty[$, we conclude\nthat $H(\\lambda) = 0$ has a unique zero in $]-1,1[$.\n\\end{proof}\n\n\\begin{remark}\nAs indicated in \\cite{elser2019learning},\nan approximate solution to \\eqref{eq:main} can be found by the Bisection method, Newton's method, or a combined\nversion. See also \\cite{burden2011analisis, robust}.\n\\end{remark}\n\n\n\\begin{theorem}\n\\label{t:hyperbola1}\nLet $u_0,v_0 \\in \\ensuremath{{X}}$, $u_{0} \\neq 0, v_{0} \\neq 0$. Then $\\proj{\\ensuremath{\\tilde{C}}_1}(u_0,v_0)$ is a singleton, and\n $$\\proj{\\ensuremath{\\tilde{C}}_1}(u_0,v_0)\n\n  =\n \n \\left\\{ \\begin{bmatrix}\n   \\frac{u_0}{1+\\lambda}\\\\\n   \\addlinespace\n   \\frac{v_0}{1-\\lambda}\n\n \\end{bmatrix}\n \\right\\}\n  $$\n in which $\\lambda$ is the unique root of $H(\\lambda)=0$ in $]-1,1[$ where\n\n  \\begin{align*}\n     H(\\lambda) =\n      \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} -1, \\text{ and }\n        \\\\\n      p =\\norm{u_0}^2 - \\norm{v_0}^2, \\quad q =\\norm{u_0}^2 + \\norm{v_0}^2.\n    \n      \\end{align*}\n\\end{theorem}\n\\begin{proof}\nApply \\cref{p:2011d4.1}.\n\\end{proof}\n\\begin{theorem}\n\\label{p:2011d4.5}\n  Let $u_0,v_0 \\in \\ensuremath{{X}}$ with either $u_0=0$ or $v_0=0$.\n\\begin{enumerate}\n\\item\n\\label{p:2011d4.5a} When $u_0 = 0,$\nwe have\n    \\begin{align}\n    \\label{f:1}\n    \\proj{\\ensuremath{\\tilde{C}}_1}(0,v_0) = \\left\\{\n    \\begin{bmatrix}\n    u\\\\\n    \\frac{v_0}{2}\n    \\end{bmatrix}\n    \\bigg|\\ \\|u\\|^2 = 2+ \\frac{\\|v_0\\|^2}{4}\n    \\right\\}.\\end{align}\n    \\item\n    \\label{p:2011d4.5b}\n    a) When $ v_0 = 0$ and $\\|u_0\\|\\geq 2\\sqrt{2}$, we have\n    \\begin{align}\n    \\label{f:2a}\n    \\proj{\\ensuremath{\\tilde{C}}_1}(u_0,0) = \\left\\{\n    \\begin{bmatrix}\n    \\frac{u_0}{2}\\\\\n    v\\end{bmatrix}\n    \\bigg| \\ \\|v\\|^2 = \\frac{\\|u_0\\|^2}{4}-2\\right\\}.\n    \\end{align}\n    b) When $v_0 = 0$ and $ 0<\\|u_0\\|< 2\\sqrt{2}$, we have\n    \\begin{align}\n    \\label{f:2b}\n    \\proj{\\ensuremath{\\tilde{C}}_1}(u_0,0) = \\left\\{\\begin{bmatrix}\n    \\sqrt{2}\\frac{u_0}{\\|u_0\\|}\\\\\n    0\\end{bmatrix}\n    \\right\\}.\\end{align}\n\n\\end{enumerate}\n\\end{theorem}\n\n\\begin{proof} By \\eqref{eq:l1}--\\eqref{eq:l3}, $(u,v)\\in \\proj{\\ensuremath{\\tilde{C}}_1}(u_{0},v_0)$ satisfies:\n\\begin{align}\n        (1+\\lambda) u = u_0 \\label{0.1}\\\\\n        (1 - \\lambda) v = v_0 \\label{0.2}\\\\\n        \\|u\\|^2 - \\|v\\|^2=2    \\label{0.3},\n\\end{align}\nfor some $\\lambda\\in\\ensuremath{\\mathbb R}$.\n\n\\noindent \\ref{p:2011d4.5a}: $u_0 = 0,v_0\\neq0$.\n\\eqref{0.1} yields\n $(1+\\lambda)u=0 \\Rightarrow$ either $1+\\lambda=0$ or $u=0$. We consider two cases.\n\n\\textbf{Case 1:}~$1+\\lambda=0$, i.e., $\\lambda =-1$.\nBy (\\ref{0.2}),\n$2v =v_0 \\Rightarrow v = \\frac{v_0}{2}$ and then\nby (\\ref{0.3}), $\\|u\\|^2-\\|\\frac{v_0}{2}\\|^2=2 \\Rightarrow \\|u\\|^2=2+\\frac{\\|v_0\\|^2}{4}$. So, $(u,v)=\\left(u,\\frac{v_0}{2}\\right)$ and\n$\\|u\\|^2=2+\\frac{\\|v_0\\|^2}{4}$. The objective function is\n\\begin{align*}\n    f(u,v) =\\|u-u_0\\|^2+\\|v-v_0\\|^2\n= \\|u\\|^2+\\norm{\\frac{v_0}{2}}^2\n    = 2+\\frac{\\|v_0\\|^2}{4}+\\frac{\\|v_0\\|^2}{4}\n    = 2+\\frac{\\|v_0\\|^2}{2}.\n\\end{align*}\n\n\\textbf{Case 2:}~$u=0$. By (\\ref{0.3}), $(0-\\|v\\|^2)=2$, which is impossible.\\\\\nCombining both cases we have the formula (\\ref{f:1}).\n\n\n\\noindent \\ref{p:2011d4.5b}: $u_0 \\neq 0$, but $v_0 = 0$, which implies $(1-\\lambda)v =0$ by (\\ref{0.2}).\nWe consider two cases:\n\n\\textbf{Case 1:} $1-\\lambda=0$, i.e., $ \\lambda = 1$.\nBy (\\ref{0.1}), $2u = u_0 \\Rightarrow u = \\frac{u_0}{2}$ and then by (\\ref{0.3}), $\\|v\\|^2 = \\frac{\\|u_0\\|^2}{4}-2 \\Rightarrow \\frac{\\|u_0\\|^2}{4}-2 \\geq 0 \\Rightarrow \\|u_0\\|^2 \\geq 8$. Thus, $(u,v)=\\left(\\frac{u_0}{2},v\\right)$ with\n$\\|v\\|^2=\\frac{\\|u_0\\|^2}{4}-2$, where the objective is\n\\begin{align*}\n    f(u,v) = \\|u-u_0\\|^2+\\|v-v_0\\|^2\n= \\frac{\\|u_0\\|^2}{4}+\\frac{\\|u_0\\|^2}{4} -2\n    = \\frac{\\|u_0\\|^2}{2} -2.\n\\end{align*}\nNote that Case 1 needs $\\|u_0\\|\\geq 2\\sqrt{2}$.\n\n\\textbf{Case 2:} $v=0$. Then $\\|u\\|^2=2 \\Rightarrow \\|u\\|=\\sqrt{2}$\nby \\eqref{0.3}. By (\\ref{0.1}), $|1+\\lambda|\\|u\\|=\\|u_0\\| \\Rightarrow |1+\\lambda| = \\frac{\\|u_0\\|}{\\sqrt{2}} \\Rightarrow 1+\\lambda = \\pm \\frac{\\|u_0\\|}{\\sqrt{2}}.$\n\\\\ \\textbf{Subcase 1:}\n$1+\\lambda = -\\frac{\\|u_0\\|}{\\sqrt{2}}$. By (\\ref{0.1}), $u = \\frac{u_0}{(-\\frac{\\|u_0\\|}{\\sqrt{2}})} = - \\frac{\\sqrt{2} u_0}{\\|u_0\\|}$. Then\n$$f(u,0)=\\|u-u_0\\|^2+\\|0-v_0\\|^2= \\norm{-\\frac{\\sqrt{2} u_0}{\\|u_0\\|}-u_0}^2\n    =\n    (\\|u_0\\|+\\sqrt{2})^2.$$\n\\\\ \\textbf{Subcase 2:}\n$1+\\lambda = \\frac{\\|u_0\\|}{\\sqrt{2}}.$\nBy \\eqref{0.1},\n$u = \\frac{\\sqrt{2} u_0}{\\|u_0\\|}$. Then\n$$f(u,0)=\\|u-u_0\\|^2+\\|0-v_0\\|^2= \\norm{u_0-\\frac{\\sqrt{2} u_0}{\\|u_0\\|}}^2\n= (\\|u_0\\|-\\sqrt{2})^2. $$\nComparing Subcase 1 and Subcase 2, we obtain $|\\|u_0\\|-\\sqrt{2}|<\\|u_0\\|+\\sqrt{2}$ because $\\|u_0\\|\\neq 0$.\nThat is, Subcase 2 gives a smaller value at $(u,0)$ with $u=\\sqrt{2}\\frac{u_0}{\\|u_0\\|}$. We need to compare it\nto Case 1 whenever it happens. Note that\n$$\\frac{\\|u_0\\|^2}{2}-2<(\\|u_0\\|-\\sqrt{2})^2\\quad\\text{whenever}\\quad\\|u_0\\|\\neq 2\\sqrt{2}.$$\nIndeed, we have\n$$\n    \\frac{\\|u_0\\|^2}{2}-2<\\|u_0\\|^2-2\\sqrt{2}\\|u_0\\|+2 \\Longleftrightarrow \\frac{\\|u_0\\|^2}{2} -2\\sqrt{2}\\|u_0\\|+4 > 0\\\\\\Longleftrightarrow\n   \n    (\\|u_0\\|-2\\sqrt{2})^2 >0,\n$$\nwhich always holds if $\\|u_0\\|\\neq 2\\sqrt{2}$. Hence, the nearest points are given by\n\\begin{align}\n\\label{smile1}\n    \\left(\\frac{u_0}{2},v\\right)\\quad\n    \\text{with}\\quad \\|v\\|^2 = \\frac{\\|u_0\\|^2}{4}-2,\\quad\\text{ when}\\quad \\|u_0\\|\\neq 2\\sqrt{2}\n\\end{align}\n\\begin{align}\n\\label{smile2}\n\\left(\\frac{u_0}{2},0\\right)\n=\\left(\\sqrt{2}\\frac{u_0}{\\|u_0\\|},0\\right), \\quad\\text{ when}\\quad \\|u_0\\|=2\\sqrt{2}.\n\\end{align}\nHowever, Case 1 occurs only when\n$\\|u_0\\|\\geq 2\\sqrt{2}$. Hence, when\n$\\|u_0\\|\\geq 2\\sqrt{2}$ the nearest points are given by\n$$\\left(\\frac{u_0}{2},v\\right)\n\\quad \\text{with} \\quad \\|v\\|^2 = \\frac{\\|u_0\\|^2}{4}-2.$$\nWhen $\\|u_0\\|<2\\sqrt{2}$, Case 1 is impossible.\nThen we only need to compare Subcase 1 and Subcase 2. Hence, the nearest point is\n$$\\left(\\sqrt{2}\\frac{u_0}{\\|u_0\\|},0\\right)\\quad\\text{when}\\quad\\|u_0\\|<2\\sqrt{2}.$$\nFinally, since $\\|u_0\\|=2\\sqrt{2} \\Rightarrow \\|v\\|^2=2-2=0$, (\\ref{smile1}) gives $\\left(\\frac{u_0}{2},\n0\\right)$. Also $\\frac{\\sqrt{2}}{\\|u_0\\|}=\\frac{1}{2}$, therefore $\\sqrt{2}\\frac{u_0}{\\|u_0\\|}=\\ensuremath{\\tfrac{1}{2}} u_0$.\nIt follows that\n$\\left(\\frac{u_0}{2},0\\right)=\\left(\\sqrt{2}\\frac{u_0}{\\|u_0\\|},0\\right)$\nwhen $\\|u_0\\|=2\\sqrt{2}$, which implies that\n(\\ref{smile2}) can be obtained from (\\ref{smile1}) when $\\|u_0\\|=2\\sqrt{2}$.\nHence formulas (\\ref{f:2a}) and (\\ref{f:2b}) hold.\n\\end{proof}\n\n\n\n\\subsection{Finding the projection $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}$}\n\n\\noindent $\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}$ can be found via $\\proj{\\ensuremath{\\tilde{C}}_1}$.\n\n\\begin{theorem}\n\\label{t:201147}\nLet $\\gamma>0, \\tilde{u}_0,\\tilde{v}_0 \\in \\ensuremath{{X}}$, and\n$\\ensuremath{\\tilde{C}}_{\\gamma} =\n      \\menge{(\\tilde{u},\\tilde{v})\\in \\ensuremath{{X}} \\times \\ensuremath{{X}}}{\\norm{\\tilde{u}}^2 -\\norm{\\tilde{v}}^2 = 2\\gamma}$. Then the following hold:\n        \n\\begin{enumerate}\n\\item \\label{t:201147.a} When $\\tilde{u}_0 \\neq 0$ and $ \\tilde{v}_0 \\neq 0$,\nwe have\n$$\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\tilde{u}_0,\\tilde{v}_0)\n=\\left\\{\\begin{bmatrix}\n   \\frac{\\tilde{u}_0}{1+\\lambda}\\\\\n   \\addlinespace\n   \\frac{\\tilde{v}_0}{1-\\lambda}\n\n \n  \\end{bmatrix}\n  \\right\\},\n  $$\nin which $\\lambda$ is the unique root of $H(\\lambda)=0$ in $\\left]-1, 1\\right[$, where\n  \\begin{align*}\n     H(\\lambda) =\n      \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} -\\gamma,~\n      p =\\norm{\\tilde{u}_0}^2 - \\norm{\\tilde{v}_0}^2,~ q =\\norm{\\tilde{u}_0}^2 + \\norm{\\tilde{v}_0}^2.\n      \\end{align*}\n\\item \\label{t:201147.b} When $\\tilde{u}_0 = 0,$\nwe have\n    \\begin{align}\n    \\label{k:1}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(0,\\tilde{v}_0) = \\left\\{\n    \\begin{bmatrix}\\tilde{u}\\\\\n    \\frac{\\tilde{v}_0}{2}\\end{bmatrix}\n    \\bigg|\\ \\|\\tilde{u}\\|^2 =\n    2\\gamma+ \\frac{\\|\\tilde{v}_0\\|^2}{4}\\right\\}.\\end{align}\n    \\item\n    \\label{t:201147.c}\n    a) When $ \\tilde{v}_0 = 0$ and $ \\|\\tilde{u}_0\\|\\geq 2\\sqrt{2\\gamma}$, we have\n    \\begin{align}\n    \\label{k:2a}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\tilde{u}_0,0) = \\left\\{\n    \\begin{bmatrix}\n    \\frac{\\tilde{u}_0}{2}\\\\\n    \\tilde{v}\n    \\end{bmatrix}\n    \\bigg| \\ \\|\\tilde{v}\n    \\|^2 = \\frac{\\|\\tilde{u}_0\\|^2}{4}-2\\gamma\\right\\}.\n    \\end{align}\n    b) When $ \\tilde{v}_0 = 0$ and\n    $0<\\|\\tilde{u}_0\\|< 2\\sqrt{2\\gamma}$, we have\n    \\begin{align}\n    \\label{k:2b}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\tilde{u}_0,0) = \\left\\{\n    \\begin{bmatrix}\n    \\sqrt{2\\gamma}\\frac{\\tilde{u}_0}{\\|\\tilde{u}_0\\|}\\\\\n    0\n    \\end{bmatrix}\n    \\right\\}.\n    \\end{align}\n\\end{enumerate}\n\\end{theorem}\n\\begin{proof}\nProposition~\\ref{p:reformulate}\\ref{i:scaling} states\n$$\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\tilde{u}_0,\\tilde{v}_0)=\\sqrt{\\gamma}\\proj{\\ensuremath{\\tilde{C}}_1}\\left(\n\\frac{\\tilde{u}_{0}}{\\sqrt{\\gamma}},\\frac{\\tilde{v}_{0}}{\\sqrt{\\gamma}}\\right).$$\nApply \\cref{t:hyperbola1} and \\cref{p:2011d4.5}.\n\\end{proof}\n\n\n\\section{Projections onto bilinear constraint set $C_{\\gamma}$ when $\\gamma>0$}\n\\label{sec:positive:gam}\n$P_{C_{\\gamma}}$ can be found via $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}$, which is the main result of this section.\n\\begin{theorem}\n\\label{t:withgamma}\nLet $\\gamma>0,~x_0,y_0 \\in X$, and $C_{\\gamma} =\n      \\menge{(x,y)\\in X \\times X}{\\scal{x}{y} = \\gamma}$. \nThen the following hold:\n\\begin{enumerate}\n    \\item \\label{i:xy:notbothe}\n    When $x_0 \\neq  \\pm y_0$, the projection is a singleton:\n   \n    \\begin{align*}\n\\proj{C_{\\gamma}}(x_0,y_0)\n          =\\left\\{\\begin{bmatrix}\n    \\frac{x_0-\\lam y_0}{1-\\lam^2} \\\\\n    \\addlinespace\n    \\frac{y_0-\\lam x_0}{1-\\lam^2}\n    \\end{bmatrix}\\right\\},\n    \\end{align*}\n    in which $\\lambda$ is the unique solution of $H(\\lambda)=0$ in $\\left]-1,1\\right[$, where\n    \\begin{align*}\n      H(\\lambda) =\n      \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} -\\gamma,~\n      p =2\\scal{x_0}{y_0},~\\text{ and } q =\\norm{x_0}^2 + \\norm{y_0}^2.\n  \\end{align*}\n   \\item\\label{i:u:zero}\n    When $x_0 =-y_0$,  the projection is a set:\n    $$\\proj{C_{\\gamma}}(x_0,-x_0) =\\left\\{\n  \\begin{bmatrix}\n  \\frac{x_{0}}{2}+\\frac{\\ensuremath{\\tilde{u}}}{\\sqrt{2}}\\\\\n  \\addlinespace\n  -\\frac{x_0}{2}+\\frac{\\ensuremath{\\tilde{u}}}{\\sqrt{2}}\n   \\end{bmatrix}\n  \\bigg|\\ \\|\\ensuremath{\\tilde{u}}\\|^2=2\\gamma+ \\frac{\\|x_0\\|^2}{2},~~ \\ensuremath{\\tilde{u}} \\in X \\right\\}.$$\n\n    \\item\\label{i:v:zero1}\n    a) When $x_0 =y_0$ and $\\|x_0\\|\\geq 2\\sqrt{\\gamma}$,\n    the projection is a set:\n    $$\\proj{C_{\\gamma}}(x_0,x_0) =  \\left\\{\n   \n   \\begin{bmatrix}\n   \\frac{x_0}{2}-\\frac{\\ensuremath{\\tilde{v}}}{\\sqrt{2}}\\\\\n   \\addlinespace\n   \\frac{x_0}{2}+\\frac{\\ensuremath{\\tilde{v}}}{\\sqrt{2}}\n   \\end{bmatrix}\n   \\bigg|\\ \\|\\ensuremath{\\tilde{v}}\\|^2=\\frac{\\|x_0\\|^2}{2}-2\\gamma,~~ \\ensuremath{\\tilde{v}} \\in X \\right\\}.$$\n  b) When $x_0 =y_0$ and $0<\\|x_0\\|< 2\\sqrt{\\gamma}$, the projection\n  is a singleton:\n    $$\\proj{C_{\\gamma}}(x_0,x_0) =  \\left\\{\\sqrt{\\gamma}\\begin{bmatrix}\\frac{x_{0}}{\\|x_{0}\\|}\\\\\n    \\addlinespace\n    \\frac{x_{0}}{\\|x_{0}\\|}\n    \\end{bmatrix}\\right\\}.\n    $$\n   \n \n   \\end{enumerate}\n\\end{theorem}\n\\begin{proof} Proposition~\\ref{p:reformulate}\\ref{i:rotate} states\n\\begin{equation}\\label{e:pformula}\n\\proj{C_{\\gamma}}(x_0,y_0) = A_{\\frac{\\pi}{4}}\n\n\n\n \n \n \n\n  \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}A_{-\\frac{\\pi}{4}}(x_{0},y_{0}).\n \n    \\end{equation}\nIt suffices to apply \\cref{t:201147}. Indeed, with\n\\begin{equation}\\label{e:point}\n\\begin{bmatrix}\n\\tilde{u}_{0}\\\\\n\\tilde{v}_{0}\n\\end{bmatrix} =A_{-\\pi/4}\\begin{bmatrix}\nx_{0}\\\\\ny_{0}\n\\end{bmatrix}=\\begin{bmatrix}\n\\frac{x_{0}+y_{0}}{\\sqrt{2}}\\\\\n\\addlinespace\n\\frac{-x_{0}+y_{0}}{\\sqrt{2}}\n\\end{bmatrix},\n\\end{equation}\nusing \\eqref{e:pformula} and \\eqref{e:point} Theorem~\\ref{t:201147} gives:\n\\bigskip\n\n\n\\ref{i:xy:notbothe}:  When $x_{0}\\neq \\pm y_{0}$, we have $\\ensuremath{\\tilde{u}}_{0}\\neq 0, \\ensuremath{\\tilde{v}}_{0}\\neq 0$.\nBy \\cref{t:201147}\\cref{t:201147.a}, we get\n$\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\ensuremath{\\tilde{u}}_0,\\ensuremath{\\tilde{v}}_0)= [\\frac{\\ensuremath{\\tilde{u}}_0}{(1+\\lam)},\\frac{\\ensuremath{\\tilde{v}}_0}{(1-\\lam)}]^T$, so\n\\begin{align*}\n\\proj{C_{\\gamma}}(x_0,y_0) &= \\left[ {\\begin{array}{cc}\n   \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} & -\\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}}\\\\\n   \\addlinespace\n   \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} & \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} \\\\\n  \\end{array} } \\right]\\left[ {\\begin{array}{cc}\n   \\frac{1}{1+\\lambda}\\ensuremath{\\operatorname{Id}} & 0\\\\\n   \\addlinespace\n   0 & \\frac{1}{1-\\lambda}\\ensuremath{\\operatorname{Id}} \\\\\n  \\end{array} } \\right]\\left[ {\\begin{array}{cc}\n   \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} & \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}}\\\\\n   \\addlinespace\n   -\\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} & \\frac{1}{\\sqrt{2}}\\ensuremath{\\operatorname{Id}} \\\\\n  \\end{array} } \\right]\\begin{bmatrix}\n           x_0 \\\\\n          y_0 \\end{bmatrix}\\\\&=  \\left[ {\\begin{array}{cc}\n   \\frac{1}{1-\\lambda^2}\\ensuremath{\\operatorname{Id}} & \\frac{-\\lambda}{1-\\lambda^2}\\ensuremath{\\operatorname{Id}}\\\\\n   \\addlinespace\n   \\frac{-\\lambda}{1-\\lambda^2}\\ensuremath{\\operatorname{Id}} & \\frac{1}{1-\\lambda^2}\\ensuremath{\\operatorname{Id}} \\\\\n  \\end{array} } \\right]\\left[ {\\begin{array}{cc}\n   x_0\\\\\n   y_0\n  \\end{array} } \\right].\n \\end{align*}\n Also,\n $\\|p\\|=\\|\\ensuremath{\\tilde{u}}_{0}\\|^2-\\|\\ensuremath{\\tilde{v}}_{0}\\|^2=2\\scal{x_{0}}{y_{0}},$ and\n $\\|q\\|=\\|\\ensuremath{\\tilde{u}}_{0}\\|^2+\\|\\ensuremath{\\tilde{v}}_{0}\\|^2=\\|x_{0}\\|^2+\\|y_{0}\\|^2$.\n\n\n\n  \\ref{i:u:zero}:\n  For $x_0=-y_0$, we have $\\ensuremath{\\tilde{u}}_{0}=0, \\ensuremath{\\tilde{v}}_{0}=-\\sqrt{2}x_{0}$, so\n  \\begin{align*}\n      \\proj{C_{\\gamma}}(x_0,-x_0)&=A_{\\frac{\\pi}{4}}\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}\n      }(0,-\\sqrt{2}x_0)\\\\\n  &=\\left\\{A_{\\frac{\\pi}{4}}\\begin{bmatrix}\n  \\ensuremath{\\tilde{u}}\\\\\n  \\addlinespace\n  \\frac{-x_0}{\\sqrt{2}}\n  \\end{bmatrix}\\bigg|\\|\\ensuremath{\\tilde{u}}\\|^2=2\\gamma+\\frac{\\|-\\sqrt{2}x_0\\|^2}{4}=2\\gamma+\\frac{\\|x_0\\|^2}{2}, \\ensuremath{\\tilde{u}}\\in X\\right\\}.\n  \\end{align*}\n\n  \\ref{i:v:zero1}:\n  When $x_0\\neq -y_0$, and $x_0=y_0$, we have $x_0\\neq 0, (x_0,y_0)=(x_0,x_0)$, so that\n  $\\ensuremath{\\tilde{v}}_{0}=0$ and\n  $\\ensuremath{\\tilde{u}}_0 = \\frac{x_0+y_0}{\\sqrt{2}}=\\sqrt{2}x_0.$  Then\n  $\\|\\ensuremath{\\tilde{u}}_0\\|\\geq 2\\sqrt{2\\gamma} \\Leftrightarrow \\|x_0\\|\\geq 2\\sqrt{\\gamma}.$\n\n  a) When $\\|x_{0}\\|\\geq 2\\sqrt{\\gamma}$, we have\n  \\begin{align*}\n    \\proj{C_{\\gamma}}(x_0,x_0)&=A_{\\frac{\\pi}{4}}\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\sqrt{2}x_0,0)\\\\\n    &=A_{\\frac{\\pi}{4}}\\left\\{\\begin{bmatrix}\n  \\frac{x_0}{\\sqrt{2}}\\\\\n  \\addlinespace\n  \\ensuremath{\\tilde{v}}\n  \\end{bmatrix}\\bigg|\\ \\|\\ensuremath{\\tilde{v}}\\|^2=\\frac{\\|\\sqrt{2}x_0\\|^2}{4}-2\\gamma=\\frac{\\|x_0\\|^2}{2}-2\\gamma,\n  \\ensuremath{\\tilde{v}}\\in X \\right\\}.\n  \\end{align*}\n\n  b) When $0<\\|x_0\\|<2\\sqrt{\\gamma}$, we have $\\|\\ensuremath{\\tilde{u}}_0\\|<2\\sqrt{2\\gamma}$, so\n  \\begin{align*}\n      \\proj{C_{\\gamma}}(x_0,x_0)&=A_{\\frac{\\pi}{4}}\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}\n      (\\sqrt{2}x_0,0)=A_{\\frac{\\pi}{4}}\\begin{bmatrix}\n  \\sqrt{2\\gamma} \\frac{\\sqrt{2} x_0}{\\|\\sqrt{2}x_0\\|}\\\\\n  \\addlinespace\n  0\n  \\end{bmatrix}\n  =A_{\\frac{\\pi}{4}}\n  \\begin{bmatrix}\n  \\sqrt{2\\gamma} \\frac{x_0}{\\|x_0\\|}\\\\\n  \\addlinespace\n  0\n  \\end{bmatrix}.\n  \\end{align*}\n\n\\end{proof}\n\n\\section{Projections onto hyperbola $\\ensuremath{\\tilde{C}}_{\\gamma}$ and bilinear constraint set $C_{\\gamma}$ when $\\gamma<0$}\n\\label{sec:negative:gam}\nArmed with the results in Sections~\\ref{section:hyperbola} and ~\\ref{sec:positive:gam},\nwe can study $P_{\\ensuremath{\\tilde{C}}_{\\gamma}}$ and\n$P_{C_{\\gamma}}$ when $\\gamma<0$.\nDefine\n$$T_{1}: X\\times X\\rightarrow X\\times X: (x,y)\\mapsto (y,x),\\text{ and }$$\n$$T_{2}:X\\times X\\rightarrow X\\times X: (x,y)\\mapsto (x,-y).$$\n\\begin{theorem}\n\\label{t:20114711}\nLet $\\gamma<0, \\ensuremath{\\tilde{u}}_0,\\ensuremath{\\tilde{v}}_0 \\in \\ensuremath{{X}}$, and $\\ensuremath{\\tilde{C}}_{\\gamma} =\n      \\menge{(u,v)\\in \\ensuremath{{X}} \\times \\ensuremath{{X}}}{\\norm{u}^2 -\\norm{v}^2 = 2\\gamma}$.\nThen the following hold:\n\\begin{enumerate}\n\\item \\label{t:20114711.a} When $\\ensuremath{\\tilde{u}}_0 \\neq 0, \\ensuremath{\\tilde{v}}_0 \\neq 0$, we have\n$$\\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\ensuremath{\\tilde{u}}_0,\\ensuremath{\\tilde{v}}_0)\n  =\n \n  \\left\\{\\begin{bmatrix}\n   \\frac{\\ensuremath{\\tilde{u}}_0}{1-\\lambda}\\\\\n   \\addlinespace\n   \\frac{\\ensuremath{\\tilde{v}}_0}{1+\\lambda}\n \n  \\end{bmatrix}\n  \\right\\}\n  $$\n in which $\\lambda$ is the unique root of $H(\\lambda)=0$ in $\\left]-1,1\\right[$,\n  where \\begin{align*}\n     H(\\lambda) =\n      \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} +\\gamma,~\n      p =\\norm{\\ensuremath{\\tilde{v}}_0}^2 - \\norm{\\ensuremath{\\tilde{u}}_0}^2,~\\text{ and } q =\\norm{\\ensuremath{\\tilde{u}}_0}^2 + \\norm{\\ensuremath{\\tilde{v}}_0}^2.\n      \\end{align*}\n      \\item \\label{t:20114711.c} When $\\ensuremath{\\tilde{v}}_0 = 0$, we have\n    \\begin{align}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(\\ensuremath{\\tilde{u}}_0,0) =\n    \\left\\{\n    \\begin{bmatrix}\n    \\frac{\\ensuremath{\\tilde{u}}_0}{2}\\\\\n    \\ensuremath{\\tilde{v}}\n    \\end{bmatrix}\n    \\bigg|\\ \\|\\ensuremath{\\tilde{v}}\\|^2 =  \\frac{\\|\\ensuremath{\\tilde{u}}_0\\|^2}{4}-2\\gamma\\right\\}.\n    \\end{align}\n\n      \\item \\label{t:20114711.b}\n    a) When $\\ensuremath{\\tilde{u}}_0 = 0$ and $ \\|\\ensuremath{\\tilde{v}}_0\\|\\geq 2\\sqrt{2(-\\gamma)}$, we have\n    \\begin{align}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(0,\\ensuremath{\\tilde{v}}_0) = \\left\\{\n    \\begin{bmatrix}\n    \\ensuremath{\\tilde{u}}\\\\\n    \\frac{\\ensuremath{\\tilde{v}}_0}{2}\n    \\end{bmatrix}\n   \\bigg|\\\n    \\|\\ensuremath{\\tilde{u}}\\|^2 = \\frac{\\|\\ensuremath{\\tilde{v}}_0\\|^2}{4}+2\\gamma\\right\\}.\n    \\end{align}\n    b) When $\\ensuremath{\\tilde{u}}_0 = 0$ and $0<\\|\\ensuremath{\\tilde{v}}_0\\|< 2\\sqrt{2(-\\gamma)}$, we have\n    \\begin{align}\n    \\proj{\\ensuremath{\\tilde{C}}_{\\gamma}}(0,\\ensuremath{\\tilde{v}}_0) = \\left\\{\\begin{bmatrix}\n    0\\\\\n    \\sqrt{2(-\\gamma)}\\frac{\\ensuremath{\\tilde{v}}_0}{\\|\\ensuremath{\\tilde{v}}_0\\|}\\end{bmatrix}\n    \\right\\}.\n    \\end{align}\n\\end{enumerate}\n\\end{theorem}\n\\begin{proof}\nSince\n$$\\minimize ~ \\|\\ensuremath{\\tilde{u}}-\\ensuremath{\\tilde{u}}_{0}\\|^2+\\|\\ensuremath{\\tilde{v}}-\\ensuremath{\\tilde{v}}_{0}\\|^2 \\quad \\text{subject to}\\quad \\norm{\\ensuremath{\\tilde{u}}}^2 -\\norm{\\ensuremath{\\tilde{v}}}^2= 2\\gamma$$\nis equivalent to\n$$\\minimize ~ \\|\\ensuremath{\\tilde{v}}-\\ensuremath{\\tilde{v}}_{0}\\|^2+\n\\|\\ensuremath{\\tilde{u}}-\\ensuremath{\\tilde{u}}_{0}\\|^2 \\quad \\text{subject to}\\quad \\norm{\\ensuremath{\\tilde{v}}}^2 -\\norm{\\ensuremath{\\tilde{u}}}^2= 2(-\\gamma),$$\nwe have\n$P_{\\ensuremath{\\tilde{C}}_{\\gamma}}=T_{1}P_{\\ensuremath{\\tilde{C}}_{-\\gamma}}T_{1}.$ It suffices to apply\n\\cref{t:201147}.\n\\end{proof}\n\n\\begin{theorem}\n\\label{t:negativegamma}\nLet $\\gamma<0,~x_0,y_0 \\in X$, and $C_{\\gamma} =\n      \\menge{(x,y)\\in X \\times X}{\\scal{x}{y} = \\gamma}$. \nThen the following hold:\n\\begin{enumerate}\n    \\item\n    When $x_0 \\neq  \\pm y_0$, the projection is a singleton:\n   \n    \\begin{align*}\n\\proj{C_{\\gamma}}(x_0,y_0)\n          =\n          \\left\\{\\begin{bmatrix}\n    \\frac{x_0+\\lam y_0}{1-\\lam^2} \\\\\n    \\addlinespace\n    \\frac{y_0+\\lam x_0}{1-\\lam^2}\n    \\end{bmatrix}\n    \\right\\}\n     \\end{align*}\n    in which  $\\lambda$ is the unique solution of $H(\\lambda)=0$ in\n    $\\left]-1,1\\right[$, where\n    \\begin{align*}\n      H(\\lambda) =\n      \\frac{(\\lambda^2 + 1 ) p - 2\\lambda q}{2(1-\\lambda^2)^2} +\\gamma,~\n      p =-2\\scal{x_0}{y_0},~\\text{ and } q =\\norm{x_0}^2 + \\norm{y_0}^2.\n  \\end{align*}\n   \\item\n    When $x_0 =y_0$,  the projection is a set:\n    $$\\proj{C_{\\gamma}}(x_0,x_0) =\\left\\{\n    \\begin{bmatrix}\n    \\frac{x_{0}}{2}+\\frac{\\ensuremath{\\tilde{u}}}{\\sqrt{2}}\\\\\n    \\addlinespace\n    \\frac{x_{0}}{2}-\\frac{\\ensuremath{\\tilde{u}}}{\\sqrt{2}}\n    \\end{bmatrix}\n    \\bigg|\\ \\|\\ensuremath{\\tilde{u}}\\|^2=-2\\gamma+ \\frac{\\|x_0\\|^2}{2},~~ \\ensuremath{\\tilde{u}} \\in X\\right\\}.$$\n\n    \\item\n    a) When $x_0 =-y_0$ and $\\|x_0\\|\\geq 2\\sqrt{-\\gamma}$,\n    the projection is a set:\n    $$\\proj{C_{\\gamma}}(x_0,-x_0) =  \\left\\{\n    \\begin{bmatrix}\n    \\frac{x_{0}}{2}-\\frac{\\ensuremath{\\tilde{v}}}{\\sqrt{2}}\\\\\n    \\addlinespace\n    \\frac{-x_{0}}{2}-\\frac{\\ensuremath{\\tilde{v}}}{\\sqrt{2}}\n    \\end{bmatrix}\n    \\bigg|\\ \\|\\ensuremath{\\tilde{v}}\\|^2=\\frac{\\|x_0\\|^2}{2}+2\\gamma,\n    ~~ \\ensuremath{\\tilde{v}} \\in X\\right\\}.$$\n  b) When $x_0 =-y_0$ and $0<\\|x_0\\|< 2\\sqrt{-\\gamma}$, the projection\n  is a singleton:\n    $$\\proj{C_1}(x_0,-x_0) =  \\left\\{\n    \\sqrt{-\\gamma}\n    \\begin{bmatrix}\n    \\frac{x_{0}}{\\|x_{0}\\|}\\\\\n    \\addlinespace\n    \\frac{-x_{0}}{\\|x_{0}\\|}\n    \\end{bmatrix}\n    \\right\\}.$$\n \n   \\end{enumerate}\n\\end{theorem}\n\\begin{proof} Since\n$$\\minimize ~ \\|x-x_{0}\\|^2+\\|y-y_{0}\\|^2 \\quad \\text{subject to}\\quad \\scal{x}{y}= \\gamma$$\nis equivalent to\n$$\\minimize ~ \\|x-x_{0}\\|^2+\\|z-(-y_{0})\\|^2 \\quad \\text{subject to}\\quad \\scal{x}{z}= -\\gamma,$$\nwe have\n$P_{C_{\\gamma}}=T_{2}P_{C_{-\\gamma}}T_{2}$. It remains to apply Theorem~\\ref{t:withgamma}.\n\\end{proof}\n\n\n\n\n\n\n\n\n\n\n\\section*{Acknowledgments}\nHHB and XW were supported by NSERC Discovery grants. MKL was supported by SERB-UBC fellowship and HHB and XW's NSERC Discovery grants. This work was presented at the SIAM Conference on Optimization in July 20--23, 2021 by MKL.\n\n", "meta": {"timestamp": "2021-12-07T02:03:49", "yymm": "2112", "arxiv_id": "2112.02181", "language": "en", "url": "https://arxiv.org/abs/2112.02181", "abstract": "Sets of bilinear constraints are important in various machine learning models. Mathematically, they are hyperbolas in a product space. In this paper, we give a complete formula for projections onto sets of bilinear constraints or hyperbolas in a general Hilbert space.", "subjects": "Optimization and Control (math.OC); Functional Analysis (math.FA); Numerical Analysis (math.NA)", "title": "Projections onto hyperbolas or bilinear constraint sets in Hilbert spaces", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9820137863531805, "lm_q2_score": 0.861538211208597, "lm_q1q2_score": 0.8460424008769004}}
{"url": "https://arxiv.org/abs/1610.01568", "title": "The ratio of domination and independent domination numbers on trees", "abstract": "Let $\\gamma(G)$ and $i(G)$ be the domination number and the independent domination number of $G$, respectively. In 1977, Hedetniemi and Mitchell began with the comparison of of $i(G)$ and $\\gamma(G)$ and recently Rad and Volkmann posted a conjecture that $i(G)/ \\gamma(G) \\leq \\Delta(G)/2$, where $\\Delta(G)$ is the maximum degree of $G$. In this work, we prove the conjecture for trees and provide the graph achieved the sharp bound.", "text": "\\section{Introduction}\nThroughout this paper $G = (V, E)$ is a   simple undirected graph with vertex set $  V(G)$ and edge set $ E(G)$.    For  $v\\in V(G)$,  $N_G(v)=\\{ w\\in V(G):  vw\\in E(G)\\}$ is \nthe  open neighborhood of $v$ and $N_G[v]=N_G(v) \\cup \\{v\\}$ is \nthe  closed neighborhood of $v$ in $G$.  If $N_G(v) = \\phi$,   $v$ is called an isolated vertex.    For $S\\subseteq V(G)$, $N_G(S)$ is the open neighborhood of $S$,  $N_G[S]=N_G(S)\\cup S$ is the  closed neighborhood of $S$ and  $G-S$ is a subgraph induced by $V(G)-S$.    A graph $F$ is   a forest if it has no cycles.  Specially,  $F$ is a tree if it contains only one component. A double star is a tree with exactly two vertices of  degree greater than 1. In paticular, if the two vertices have same degree, then it is called a balanced double star. \nThe line graph $L(G)$ of  a connected graph is a graph such that each vertex of $L(G)$ represents an edge of $G$  and two vertices of $L(G)$ are adjacent if and only if their corresponding edges share a common endpoint in $G$.\n\n \nIt is known that a vertex set $D \\subset V(G)$  is a  dominating set if every vertex of $V(G)-D$ is adjacent to some vertices of $D$. The minimum cardinality of a dominating set is called the domination number, denoted by $\\gamma(G)$.  Similarly,\na vertex set $I \\subset V(G)$  is  an independent dominating set if $I$ is both an independent set and a dominating set in $G$, where an independent set  is a set of vertices in a graph such that  no two of which are adjacent. The minimum cardinality of  an independent dominating set is called the  independent domination number, denoted by $ i(G)$. Currently, lots of  work relating domination number and independent domination number have been studied, referred to surveys \\cite{2,4}.\n\n In 1977,  S. Hedetniemi and S. Mitchell \\cite {1977} showed that for any tree T,\n $\\frac{ i(L(T))} { \\gamma (L(T))} =1$, where $ L(T)$ is the line graph of T.\nBecause any line graph is a $K_{1,3}$-free graph, \n R. B. Allan and R. Laskar \\cite {1978} extended the previous result  in 1978 and obtained that if a graph  does not have an induced subgraph isomorphic to $K_{1,3}$, then $  i(G)/ \\gamma(G) = 1$.\nRecently, Goddard et al.\\cite{3} considered the ratio $i(G)/ \\gamma(G) $ for regular graphs and proved that $i(G)/ \\gamma(G) \\leq  3/2$ for cubic graphs. In 2013, Southey and Henning \\cite{6} improved the previous result to $i(G)/ \\gamma(G)  \\leq 4/3$ for connected cubic graphs except for $K_{3,3}$.  \nDuring the same year, Rad and Volkmann \\cite{5}  got an upper bound of $i(G)/\\gamma(G)$ for a graph $G$ and prosposed the conjecture.\n \\vskip 2mm {\\bf Theorem 1}\\emph{ (Rad and Volkmann \\cite{5}) \nLet $G$ be a graph, then\n$$ \\frac{i(G)}{\\gamma(G)} \\leq\n \\left\\{ \\begin{array}{rcl}\n   \\frac{\\Delta(G)}{2},\\;\\;\\;\\;\\;\\;\\;\\; \\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;&& \\mbox{if } 3 \\leq \\Delta(G) \\leq 5,\n  \\\\\\Delta(G) -3 +\\frac{2}{\\Delta(G) - 1},&&   \\mbox{if } \\Delta(G) \\geq 6.\n\\end{array}\\right.$$\n}\n{\\bf Conjecture 2} \\emph{(Rad and Volkmann \\cite{5}) \nLet $G$  be a graph with $\\Delta(G) \\geq 3$, then $i(G)/\\gamma(G) \\leq \\Delta(G)/2$.\n}\n\nIn 2014, Furuta et al.\\cite{7} showed that $i(G)/ \\gamma(G)  \\leq \\Delta(G) - 2 \\sqrt{\\Delta(G)} +2$  for  a graph $G$ and gave the graph achieved the new bound.  However,\nwhen $\\Delta(G)$ is big enough, then $ \\Delta(G) - 2 \\sqrt{\\Delta(G)} +2 > \\Delta(G) / 2.$ Now there is a natural question that\\\\\n{\\it Q:  Is there  other class of graphs,  which has an affirmative answer for Conjecture 2? }\n\n\nMotivated by Conjecture 2 and the above question, we prove that Conjeture 2 is true for the tree and provide the graph $G $, which  attains the sharp bound $\\Delta(G)/2$.\n\\begin{figure}[thb]\n\\center\n\\begin{tikzpicture}\n\\tikzstyle {every node} =[fill=black!60,circle,inner sep=0.5pt,text=white] \\path \n      (2,0) node(0) {$w_1$} \n      (4,0) node(1) {$w_2$}\n      (1,1.5) node(2) {$v_1$}\n      (1,0.5) node(3) {$v_2$}\n      (1, -0.5)node(4){$*$}\n      (1,-1.5) node (5) {$v_s$}\n      (5,1.5) node(6)  {$u_1$}\n      (5,0.5) node (7) {$u_2$}\n      (5,-0.5) node (8) {$*$}  \n      (5,-1.5) node (9) {$u_s$}   \n       ;\n\\draw[black,line width=1pt]  (0)--(1) (0)--(2) (0)--(3)  (0)--(4) (0)--(5)  \n                                       (1)--(6) (1)--(7) (1)--(8) (1)--(9)      \n;\n\\end{tikzpicture}\n\\caption{A balanced double star}\n\\label{fig:cor}\n\\end{figure}\n\n \\vskip 2mm {\\bf Theorem 3} \\emph{\nLet $G$ be a forest, then $$ \\frac{i(G)}{\\gamma(G)} \\leq\n \\left\\{ \\begin{array}{rcl}\n1,\\;\\;\\; && \\mbox{if } \\Delta(G) \\leq 2,\n  \\\\\\frac{\\Delta(G)}{2},&&   \\mbox{if } \\Delta(G) \\geq 3,\n\\end{array}\\right.$$ and the equalities hold if either $\\Delta(G) \\leq 2$ or each component of $G$ is a balanced double star(see figure 1).\n}\n\nAs an immediate consequence of Theorem 3, we obtain that \n\n \\vskip 2mm {\\bf Theorem 4} \\emph{\nLet $G$ be a tree, then $$ \\frac{i(G)}{\\gamma(G)} \\leq\n \\left\\{ \\begin{array}{rcl}\n1,\\;\\;\\; && \\mbox{if } \\Delta(G) \\leq 2,\n  \\\\\\frac{\\Delta(G)}{2},&&   \\mbox{if } \\Delta(G) \\geq 3,\n\\end{array}\\right.$$ and the equalities hold if either $\\Delta(G) \\leq 2$ or $G$ is a balanced double star(see figure 1).\n}\n\n\\section{Proof of Theorem 3}\n\nIn this section, we will prove Theorem 3 and start with an interesting lemma.\n\n \\vskip 2mm {\\bf Lemma 1} \\emph{\nLet $r_1, r_2,r_3,r_4,t$ be positive numbers with $\\frac{r_1}{r_2} \\leq t$ and $\\frac{r_3}{r_4} \\leq t$. Then $\\frac{r_1+r_3}{r_2+r_4} \\leq t$.  \n}\n\nSince $r_1 \\leq r_2 t, r_3 \\leq r_4 t$, we  replace $r_1$, $r_3$ and obtain that Lemma 1 is true. Next we will give the main proof of this note.\n\n\n\n{\\bf Proof of Theorem 3.}\nFor $\\Delta(G) \\leq 1$, $G$ contains only isolated vertices or edges and $i(G) = \\gamma(G)$, that is,  $i(G)/ \\gamma(G)  =1$.  Next, we will consider the case of $\\Delta(G) \\geq 2$ and\n  begin with the case that  the forest $G$ contains only one component, that is,  $G$ is a tree.\n\nLet $D$ be a minimum dominating set of $G$. Then $G[D]$ is also \na forest. We build $\\{G_i\\}, \\{x_i\\}$ with $i \\geq 1$ as follows:  Let $G_1 = G[D]$ and $ x_1 \\in V(G_1)$ with $d_{G_1}(x_1) = 0$ or $1$; For $i \\geq 2$, if $V(G_i-N_{G_{i}}[x_{i}]) = \\phi$, then stop and set $i = k$. Otherwise, let $G_{i} = G_{i-1}-N_{G_{i-1}}[x_{i-1}]$ and $x_{i} \\in V(G_i)$ with $d_{G_i}(x_i) =0 $ or $ 1$.  \n\nSet $X = \\{x_1,x_2,...,x_k\\}$. Then $X$ is an independent dominating set of $G[D]$\nand  $\\{N_{G_i}[x_i], 1 \\leq i \\leq k\\}$ is a   partition of $D$, that is, $\\sum_{1 \\leq i \\leq k} (d_{G_i}(x_i) +1) = |D| = \\gamma(G)$. Choose $I \\subset V(G) - D$ such that \n  $X \\cup I$ is an independent dominating set of $G$, that is,  \n$\ni(G) \\leq |X|+|I| = k+ |I|.\n$\nSince $D$ is a dominating set of $G$, then $I = \\cup_{v \\in {D-X}}(N_G(v) \\cap I) = \\cup_{1 \\leq i \\leq k} (\\cup_{v \\in {N_{G_i}(x_i)}}(N_G(v) \\cap I))$.\nBy the choice of $x_i$, for $1 \\leq i \\leq k$ and $v \\in N_{G_i}(x_i)$, we have $d_{G_i}(x_i) \\leq d_{G_i}(v)$. Thus,\n$|N_G(v) \\cap I| \\leq d_G(v) - d_{G_i}(v) \\leq \\Delta(G) - d_{G_i}(x_i)$ and\n\\begin{eqnarray}\n|I|   & \\leq &\\sum_{1 \\leq i \\leq k}( \\sum _{v \\in N_{G_i}(x_i)} |N_{G}(v) \\cap I|)  \\nonumber \\\\\n&\\leq  &  \\sum_{1 \\leq i \\leq k} (\\sum_{v \\in N_{G_i}(x_i)} (\\Delta(G) - d_{G_i}(x_i)))  \\nonumber \\\\\n&= & \\sum_{1 \\leq i \\leq k} (\\sum_{v\\in N_{G_i}(x_i)} \\Delta (G)) -  \\sum_{1 \\leq i \\leq k} ( \\sum_{v \\in N_{G_i}(x_i)} d_{G_i}(x_i))  \\nonumber\\\\\n&=& (|D| - k ) \\Delta(G) - \\sum_{1 \\leq i \\leq k}d_{G_i}(x_i)^2.   \n\\end{eqnarray}\n\nBy $(1)$  and $|D| = \\gamma(G)$, we can obtain that  \\begin{eqnarray}\ni(G) &\\leq & k+ |I| \\nonumber \\\\\n&\\leq &  k+ (|D| - k ) \\Delta(G) - \\sum_{1 \\leq i \\leq k}d_{G_i}(x_i)^2  \\nonumber \\\\\n&=& \\Delta(G)\\gamma(G) - \\sum_{1 \\leq i \\leq k } (\\Delta(G) -1 +d_{G_i}(x_i)^2), \\nonumber\n\\end{eqnarray}\nthat is, $$\n\\frac{i(G)}{\\gamma(G)} \\leq  \\Delta(G) - \\frac{ \\sum_{1 \\leq i \\leq k } (\\Delta(G) -1 +d_{G_i}(x_i)^2)}{\\gamma(G)}.  \n$$\n\nNow, it suffices to show that $ - \\frac{ \\sum_{1 \\leq i \\leq k } (\\Delta(G) -1 +d_{G_i}(x_i)^2)}{\\gamma(G)} \\leq -\\frac{\\Delta(G)}{2}$, that is,\n\\begin{eqnarray} \n\\sum_{1 \\leq i \\leq k} (\\Delta(G) - 1 + d_{G_i}(x_i)^2)  &\\geq& \\frac{1}{2} \\Delta(G) \\gamma(G)  \\nonumber \\\\&=& \\frac{1}{2} \\Delta(G) (\\sum_{1 \\leq i \\leq k} (d_{G_i}(x_i) +1)) \n\\end{eqnarray}\nBy the construction of $G_i, x_i$,   $d_{G_i}(x_i) = d_{G_i}(x_i)^2 = 0$ or $1$.\nThus, $(2)$ is the same as $(3)$ below.\n\\begin{eqnarray} \n &&\\Leftrightarrow\nk\\Delta(G) - k + \\sum_{1\\leq i \\leq k} d_{G_i}(x_i)  - \\frac{1}{2} \\Delta(G)(\\sum_{1\\leq i \\leq k} d_{G_i}(x_i) )   \\nonumber \\\\&& \\;\\;\\;\\;\\; -\\frac{1}{2} \\Delta(G) k  \\geq 0  \\nonumber\n\\\\&&\n \\Leftrightarrow\n(1- \\frac{1}{2} \\Delta(G)) ((\\sum_{1\\leq i \\leq k} d_{G_i}(x_i) ) - k) \\geq 0  \n\\end{eqnarray}\n\nFurthermore, $d_{G_i}(x_i) = 0$ or $1$ yields that   $(\\sum_{1\\leq i \\leq k} d_{G_i}(x_i) ) - k \\leq 0$.\nSince $\\Delta(G) \\geq 2$,  then\n $1- \\frac{1}{2} \\Delta(G) \\leq 0$. Thus, $(3)$ is true, that is, Theorem 3 is true for the tree.\n\n\nNext we will consider the case that $G$ has more than one component. In this case,  each component of $G$ is either  an isolated vertex or a tree, say $G_1, G_2, ..., G_s$ with an integer $s \\geq 2$. For $1 \\leq j \\leq s $, if $G_j$ is an isolated vertex, then $i(G_j)/\\gamma(G_j) = 1/1 \\leq \\Delta(G)/2$; If $G_j$ is a tree, by the above proof,   $i(G_j)/\\gamma(G_j) \\leq   \\Delta(G)/2$.  Finally,   using Lemma 1,    $i(G)/\\gamma(G) \\leq  \\Delta(G)/2$ holds for the forest. \n Furthermore, if $\\Delta(G) \\leq 2$, all  forests achieve the bound; if $\\Delta(G) \\geq 3$, the union of balanced double stars attain the bound. Thus, Theorem 3 is true.\n\n $\\hfill\\Box$\n\n\n\n\n\n\n\n\n\n\n\n\n\n", "meta": {"timestamp": "2016-10-06T02:07:46", "yymm": "1610", "arxiv_id": "1610.01568", "language": "en", "url": "https://arxiv.org/abs/1610.01568", "abstract": "Let $\\gamma(G)$ and $i(G)$ be the domination number and the independent domination number of $G$, respectively. In 1977, Hedetniemi and Mitchell began with the comparison of of $i(G)$ and $\\gamma(G)$ and recently Rad and Volkmann posted a conjecture that $i(G)/ \\gamma(G) \\leq \\Delta(G)/2$, where $\\Delta(G)$ is the maximum degree of $G$. In this work, we prove the conjecture for trees and provide the graph achieved the sharp bound.", "subjects": "Combinatorics (math.CO)", "title": "The ratio of domination and independent domination numbers on trees", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9910145717782007, "lm_q2_score": 0.8519528076067262, "lm_q1q2_score": 0.8442976468056156}}
{"url": "https://arxiv.org/abs/2203.13870", "title": "A simple mnemonic to compute sums of powers", "abstract": "We give a simple recursive formula to obtain the general sum of the first $N$ natural numbers to the $r$th power. Our method allows one to obtain the general formula for the $(r+1)$th power once one knows the general formula for the $r$th power. The method is very simple to remember owing to an analogy with differentiation and integration. Unlike previously known methods, no knowledge of additional specific constants (such as the Bernoulli numbers) is needed. This makes it particularly suitable for applications in cases when one cannot consult external references, for example mathematics competitions.", "text": "\\section{Introduction}\n\nSums of powers have fascinated mathematicians for centuries. The sum of the first $N$ natural numbers is given by the simple well-known formula\n\\begin{equation}\n    \\sum_{n=1}^N n = 1+2+\\cdots + N = \\frac{N(N+1)}{2} \\ .\n    \\label{sum r=1}\n\\end{equation}\nOn the other hand, while occasionally useful, the general formulas for the sums of higher powers are much less well-known. For example, one has\n\\begin{equation}\n    \\sum_{n=1}^N n^2 = \\frac{1}{3} N^3 + \\frac{1}{2} N^2 + \\frac{1}{6} N \\ .\n    \\label{sum r=2}\n\\end{equation}\nIn fact once a formula of the type of eq. \\eqref{sum r=1} or \\eqref{sum r=2} is given, its proof is a trivial exercise by using induction. However, \\textit{guessing} the formula in the first place is non-trivial. \n\nHere we will provide a simple-to-remember method to recursively compute formulas for sums of a generic power $r \\in \\mathbb{N}$:\n\\begin{equation}\n    \\sum_{n=1}^N n^r \\equiv S(N;r) \\ .\n    \\label{sum r}\n\\end{equation}\nIn other words, our method allows us to obtain $S(N;r+1)$ from $S(N;r)$, and therefore recursively all the $S(N;r)$. In general, $S(N;r)$ will be a polynomial in $N$ of degree $r+1$.\n\n\\section{Faulhaber's Formula}\\label{sec:faulhaber}\n\nIn the early 18th century, Jacob Bernoulli obtained a general formula for the sums of powers, now known as Faulhaber's formula,\n\\begin{equation}\n    \\sum_{n=1}^N n^r = \\frac{1}{r+1}\\sum_{j=0}^r (-1)^j {r+1 \\choose j} B_j N^{r+1-j} \\ ,\n    \\label{faulhaber}\n\\end{equation}\nwhere the $B_j$ are the Bernoulli numbers with $B_1 = -1/2$. The proof of eq. \\eqref{faulhaber} is straightforward starting from one of the forms of the Euler-Maclaurin identity \\cite{spivey}.\nWhile technically this is a solution to the problem, it is very hard to remember. Moreover, despite the apparent simplicity of eq. \\eqref{faulhaber}, the computation of the Bernoulli numbers is itself non-trivial. Formulas are either implicit or involve double summations of complicated summands. For these reasons we look for an easier-to-remember method that one can perform with pen and paper without having to look up additional references. \n\n\\section{A simple mnemonic for sums of powers}\n\nHere we illustrate the mnemonic to recursively compute sums of powers, while in the next section we give a proof of the correctness of the method. Let's consider the concrete example of the computation of $\\sum_{n=1}^N n^2$. Here we first do something \\say{illegal}, that is we differentiate the whole sum with respect to $N$, and by this we mean that we differentiate term-by-term. Doing this we obtain the sum of one power lower $\\sum_{n=1}^N n^2$, which we know:\n\\begin{equation}\n    \\frac{d}{dN} \\sum_{n=1}^N n^2 \\overset{?}{=} \\sum_{n=1}^N \\frac{d}{dn} (n^2) =2\\sum_{n=1}^N n = N(N+1) = N^2+N \\ .\n    \\label{miao}\n\\end{equation}\nThis is clearly not a valid operation, however one can formally think of it as a map between polynomials that happens to be given by a term-by-term derivative. In order to recover the original sum, one may exploit the analogy with differentiation and integration, whereby integrating reverses the derivative. Thus integrating the end result of \\eqref{miao}, one obtains\n\\begin{equation}\n    \\int_0^N N^2 +N \\,\\, dN =\\frac{1}{3} N^3 + \\frac{1}{2} N^2 \\ .\n\\end{equation}\nComparing with eq. \\eqref{sum r=2}, this is almost the correct result, lacking only the linear term $N/6$. The procedure thus works as follows:\n\\begin{itemize}\n  \\item Differentiate the sum term-by-term and substitute the formula for the lower sum of power;\n  \\item Integrate the resulting polynomial;\n  \\item Add a linear term $CN$ and fix the constant $C$ by requiring that the sum give $1$ for $N=1$ as it should. \n\\end{itemize}\nThus one is able to recursively obtain all the sums of powers for any $r$ starting from either $r=1$ or trivially $r=0$. In the next section we prove that this method indeed gives the correct result. This method can be easily remembered (\\say{differentiate, then integrate and add a linear term}).\n\nNow we compute the first few sums of powers as an example. Starting with $r=0$,\n\\begin{equation}\n    \\sum_{n=1}^N n^r=\\sum_{n=1}^N 1 = N \n\\end{equation}\ntrivially. Now we apply our procedure to the $r=1$ case. First we integrate and sum the resulting sequence:\n\\begin{equation}\n    \\frac{d}{dN} \\sum_{n=1}^N n = \\sum_{n=1}^N 1 = N \\ .\n\\end{equation}\nThen we integrate and add a linear term, finding\n\\begin{equation}\n    \\sum_{n=1}^N n = \\frac{1}{2} N^2 + CN \\ .\n\\end{equation}\nSubstituting $N=1$ we find $1/2 + C = 1$, that is $C=1/2$, so\n\\begin{equation}\n    \\sum_{n=1}^N n = \\frac{1}{2} N(N + 1) \\ ,\n\\end{equation}\nwhich is the correct formula. Now for $r=2$ we have\n\\begin{equation}\n    \\frac{d}{dN} \\sum_{n=1}^N n^2 = 2\\sum_{n=1}^N n = N^2+N \\ .\n\\end{equation}\nIntegrating and adding a linear term\n\\begin{equation}\n    \\sum_{n=1}^N n^2 = \\frac{1}{3} N^3+\\frac{1}{2} N^2 + CN \\ .\n\\end{equation}\nSubstituting $N=1$ we have $1/3+1/2 + C = 1$, that is $C=1/6$, so that\n\\begin{equation}\n    \\sum_{n=1}^N n^2 = \\frac{1}{3} N^3+\\frac{1}{2} N^2 + \\frac{1}{6}N \\ .\n\\end{equation}\nFor $r=3$, we have\n\\begin{equation}\n    \\frac{d}{dN} \\sum_{n=1}^N n^3 = 3\\sum_{n=1}^N n^2 = N^3+\\frac{3}{2} N^2 + \\frac{1}{2}N \\ .\n\\end{equation}\nAgain integrating and adding a linear term\n\\begin{equation}\n    \\sum_{n=1}^N n^3 =\\frac{1}{4}N^4+\\frac{1}{2} N^3 + \\frac{1}{4}N^2 + CN \\ .\n\\end{equation}\nFor $N=1$ we have $1/4+1/2 + 1/4 + C = 1$, that is $C=0$, so\n\\begin{equation}\n    \\sum_{n=1}^N n^3 =\\frac{1}{4}N^4+\\frac{1}{2} N^3 + \\frac{1}{4}N^2 \\ .\n\\end{equation}\nFor $r=4$,\n\\begin{equation}\n    \\frac{d}{dN} \\sum_{n=1}^N n^4 = 4\\sum_{n=1}^N n^3 = N^4+2 N^3 + N^2 \\ .\n\\end{equation}\nTherefore\n\\begin{equation}\n    \\sum_{n=1}^N n^4 =\\frac{1}{5}N^5+\\frac{1}{2} N^4 + \\frac{1}{3}N^3 + CN \\ ,\n\\end{equation}\nwhere $C=-1/30$, so that\n\\begin{equation}\n    \\sum_{n=1}^N n^4 =\\frac{1}{5}N^5+\\frac{1}{2} N^4 + \\frac{1}{3}N^3 - \\frac{1}{30}N \\ .\n\\end{equation}\nIn principle one can keep going and compute the formula for an arbitrary $r$.\n\n\\section{Proof of correctness}\n\nHere we prove that the method explained in the previous section is correct. To do so, define\n\\begin{equation}\n    S(N;r) = \\sum_{n=1}^N n^r \\ ,\n\\end{equation}\nwhere $S(N;r)$ is a polynomial in $N$. Then the statement of the previous method is essentially that\n\\begin{equation}\n    S(N;r) = CN + r \\int_0^N S(N;r-1)\\,dN \n\\end{equation}\nfor some constant $C$. To show that this is true, we compute the integral using Faulhaber's formula:\n\\begin{align*}\n    r\\int_0^N S(N;r-1)\\,dN &=r\\int_0^N \\frac{1}{r}\\sum_{j=0}^{r-1} (-1)^j {r \\choose j} B_j N^{r-j}\\,dN =\\\\\n    &=\\sum_{j=0}^{r-1} (-1)^j {r \\choose j} B_j \\frac{1}{r+1-j} N^{r+1-j}=\\\\\n    &=\\frac{1}{r+1}\\sum_{j=0}^{r-1} (-1)^j {r+1 \\choose j} B_j N^{r+1-j}=\\\\\n    &=S(N;r) - \\frac{1}{r+1} (-1)^r {r+1 \\choose r} B_r N =\\\\\n    &=S(N;r) -(-1)^r B_r N\n\\end{align*}\nThat is\n\\begin{equation}\n    S(N;r) = (-1)^r B_r N + r \\int_0^N S(N;r-1)\\,dN \\ .\n\\end{equation}\nThis is precisely the statement that we meant to prove, with an explicit value for the constant $C$. While this may be useful, we prefer to determine $C$ by setting $S(N;r)=1$ for $N=1$ as it is easier to remember.\n\n", "meta": {"timestamp": "2022-03-29T02:02:42", "yymm": "2203", "arxiv_id": "2203.13870", "language": "en", "url": "https://arxiv.org/abs/2203.13870", "abstract": "We give a simple recursive formula to obtain the general sum of the first $N$ natural numbers to the $r$th power. Our method allows one to obtain the general formula for the $(r+1)$th power once one knows the general formula for the $r$th power. The method is very simple to remember owing to an analogy with differentiation and integration. Unlike previously known methods, no knowledge of additional specific constants (such as the Bernoulli numbers) is needed. This makes it particularly suitable for applications in cases when one cannot consult external references, for example mathematics competitions.", "subjects": "General Mathematics (math.GM)", "title": "A simple mnemonic to compute sums of powers", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9884918485625379, "lm_q2_score": 0.8539127473751341, "lm_q1q2_score": 0.8440857901639618}}
{"url": "https://arxiv.org/abs/1903.05910", "title": "Efficient evaluation of noncommutative polynomials using tensor and noncommutative Waring decompositions", "abstract": "This paper analyses a Waring type decomposition of a noncommuting (NC) polynomial $p$ with respect to the goal of evaluating $p$ efficiently on tuples of matrices. Such a decomposition can reduce the number of matrix multiplications needed to evaluate a noncommutative polynomial and is valuable when a single polynomial must be evaluated on many matrix tuples.In pursuit of this goal we examine a noncommutative analog of the classical Waring problem and various related decompositions. For example, we consider a \"Waring decomposition\" in which each product of linear terms is actually a power of a single linear NC polynomial or more generally a power of a homogeneous NC polynomial. We describe how NC polynomials compare to commutative ones with regard to these decompositions, describe a method for computing the NC decompositions and compare the effect of various decompositions on the speed of evaluation of generic NC polynomials.", "text": "\\section{Introduction}\n\\ \\ \\\nThe Waring problem has had a long history since it was first proposed in 1770 by Edward Waring. Initially it was a question about integers, asking whether a natural number could be written as the sum of powers of natural numbers. Later it was extended to polynomials. It concerns the question whether a given polynomial, $f(x_1, x_2, \\dots, x_n)$, can be represented by sums of powers of polynomials, where $x_i$'s are variables which commute. In this form, the Waring problem is closely related to symmetric tensor decomposition, see section \\ref{sec:ClassicalBackground}. This problem\nwas treated successfully in \\cite{RS00} and \\cite{FOS12} and\nhas been studied extensively, as is shown, for example, in \\cite{BC11} and \\cite{GV08}. In this paper, we assume that $x_i$'s are noncommutative variables, and derive solutions for the noncommutative version of the  Waring problem. Our results can be used to efficiently evaluate noncommutative polynomials on tuples of matrices.\n\n\nPursuing the  noncommutative \\ Waring Problem \\  is in the spirit of the burgeoning area called {\\it free analysis}.\nHere one takes classical problems and works out analogs with noncommutative \\ variables,\nwhich are {\\it free} of constraints.\nThese free analogues typically have interpretations for matrix or operator variables\nand their development often impacts various areas.\n\n\nOne of the original efforts here was Voiculescu's free probability,\nwhich started by developing a notion of entropy for operator variables\nand which has a become a big area having many associations to\nrandom matrix theory, \\cite{MS17}.\nSome other directions are  free\nanalytic function theory, cf. \\cite{KVV14} and   free real algebraic geometry \\cite{BKP16}\nwith some consequences for system engineering being \\cite{HMPV09}. Our paper develops the noncommutative \\ analogue of the classical Waring problem.\n\n\n\n\n\n\n\n\n\\subsection{Problem statement}\n\\ \\ \\ \\\nWe now state a natural noncommutative version of the classical polynomial Waring problem. We shall work with functions of\n $g$ noncommutative variables\n$$x = (x_1, x_2, ..., x_g)$$\nand be interested in  powers of\n  linear functions\n$$L_s(x): = A^s_1x_1 + A^s_2x_2 + ... A^s_g x_g,$$\nwhere $s$ is an index and $A^s_i\n \\in \\mathbb{R} \\ or \\ \\mathbb{C}$ for $1 \\le i \\le g$.\n\nFor any (index) tuple $\\alpha = (\\alpha_1, \\alpha_2, ..., \\alpha_d)$, where $\\alpha_i $ for $1 \\le i \\le d$ are  integers between 1 and g, we denote\n$$ x^\\alpha = x_{\\alpha_1}x_{\\alpha_2}x_{\\alpha_3}...x_{\\alpha_d}.$$\nFor example,\nif $\\alpha = ( 1, 2, 1,3)$, then\n$x^\\alpha = x_1x_2x_1x_3$.\n\n\n\\begin{defi}\nAn homogeneous NC degree $d$   polynomial $p$ has a\n$t$-term \\df{real Waring  (resp. complex Waring)\ndecomposition}\nprovided that\n$p(x)$ can be written\nas the sum of $t$ terms of  the $d^{th}$-power of  linear functions of $x$, i.e.,\n\\begin{equation} \\label{Waring_rep}\np(x)=\\sum_{s=1}^t [ A^s_1x_1 + A^s_2x_2 + ... A^s_gx_g]^d\n\\end{equation}\nwith real (resp. complex) numbers $A_j^s$.\n\\qed \\end{defi}\n\n\n\\def{\\mathbb C}{{\\mathbb C}}\n\n\\noindent \\df{NC Waring Problem: }\n\n{\\bf Determine if a noncommutative homogeneous degree $d$ polynomial $p$ has a $t$-term Waring decomposition.}\n\n\\bigskip\n\nIn this paper we will reduce this problem to the classical commutative variable Waring problem, thereby effectively solving it over ${\\mathbb C}$.\n\nWe also examine the\n\n\\bigskip\n\n \\noindent\n \\df{General NC Waring Problem:}\n\n {\\bf Determine if a homogeneous noncommutative polynomial is the sum of $d^{th}$ powers of homogeneous polynomials\nof degree $\\delta$.}\n\n\\bigskip\n\nIn a similar spirit, we reduce the NC general Waring problem to a classical Waring problem, but in more variables.\n\n\\subsection{Background on the Waring Problem}\n\\ \\ \\ \\\nThe classical commutative versions of these problems are well summarized in \\cite{FOS12}.\n\\bigskip\n\n\\subsubsection{Classical  Waring Problem.}\n\\label{sec:ClassicalBackground}\n\\ \\ \\ \\\nAccording to Theorem 2.2 of \\cite{OO12}, we have the following theorem for the  classical Waring decomposition of linear terms:\n\n\\begin{thm} \\label{classical_result}\n A homogeneous polynomial of degree $d$ in $g$ variables can always be expressed as the sum of powers of linear forms with complex coefficients. Moreover, for a general homogeneous polynomial, the number of linear forms needed is $\\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil$, except for\n \\begin{itemize}\n     \\item $d = 2$, where $g$ terms are needed\n     \\item $(d,g) = (3,5),(4,3),(4,4),(4,5)$ where $\\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil + 1$ terms are needed.\n \\end{itemize}\n\\end{thm}\n\n\\noindent\n{\\bf Waring vs Tensor Decomposition.}\n\\ \\ \\ \\\nIt is well known that the classical polynomial Waring problem is equivalent to symmetric tensor decomposition. Let $T \\in (\\mathbb C^{g})^{\\otimes d}$ be a symmetric tensor, i.e. a symmetric multiidexed array, with entries $T_\\alpha \\in \\mathbb C$ where $\\alpha= (\\alpha_1, \\dots, \\alpha_d)$ is a $d$-tuple of integers between $1$ and $g$. We may associate $T$ to a homogeneous degree d polynomial $p_T (x)$ in the commutative variables $z=(z_1, \\dots, z_g)$ by setting\n\\[\np_T (z)= \\sum_{|\\alpha|=d} T_\\alpha z^\\alpha.\n\\]\n\nSuppose $T$ has rank $r$ symmetric tensor decomposition\n\\[\nT= \\sum_{s=1}^r A^s \\otimes \\cdots \\otimes A^s \\quad \\quad \\mathrm{where \\ } d \\mathrm{\\ copies \\ of \\ } A^s \\mathrm{\\ appear \\ in \\ each \\ tensor \\ product.}\n\\]\nHere $A^s=(A_1^s, \\dots, A_g^s) \\in \\mathbb C^{g}$ for each $s$. Then it is straightforward to check that\n\\[\np_T (z) = \\sum_{s=1}^r \\left( \\sum_{i=1}^g A_i^s z_i \\right)^d.\n\\]\nThat is, a rank $r$ symmetric tensor decomposition of $T$ corresponds to a rank $r$ Waring decomposition for $p_T (z)$. By reversing this correspondence one sees that a rank $r$ Waring decomposition for a homogeneous polynomial gives a rank $r$ symmetric tensor decomposition for the associated symmetric tensor.\n\n\\noindent\n\\subsubsection{Classical General Waring Problem.}\n\\ \\ \\ \\\nThe classical commutative Waring problem can be generalized from representation by powers of linear functions to powers of any degree homogeneous polynomials. The generalized classical Waring  problem has also been well studied.\nAccording to Theorem 4 in \\cite{FOS12},\nthere is an upper bound for the number of terms needed for\nsuch  problems:\n\n\\begin{thm}\nA general homogeneous polynomial of degree $\\delta d$ in $g$ variables, where $d \\ge 2$, can be expressed as a sum of at most\n$d^{g - 1}$ $d^{th}$ powers of degree $\\delta$ homogeneous\ncomplex coefficient polynomials. Moreover, for a fixed $g$, this bound is sharp for all sufficiently large $\\delta$.\n\\end{thm}\n\n\\subsection{An easily stated result}\n\\ \\ \\ \\\nBefore stating a result we need a definition.\nDefine an \\df{indicator function} on\n an index $d$-tuple $\\alpha=(\\alpha_1, \\dots, \\alpha_d)$ by first defining\n\\index{$\\mathbbm{1}_j^{\\alpha_i}$ }\n$$\\mathbbm{1}_j^{\\alpha_i}  =\n\\begin{cases}\n1 & \\textrm{if }  \\alpha_i = j \\\\\n0 & \\textrm{if }  \\alpha_i \\neq j\n\\end{cases}.\n$$\nThen the indicator function $\\mathbbm{1}_j^\\alpha$ which gives the number of $j$'s appearing in $\\alpha$ is\n\\index{$\\mathbbm{1}_j^{\\alpha}$ }\n$$\n\\mathbbm{1}_j^\\alpha := \\sum_{i=1}^d \\mathbbm{1}_j^{\\alpha_i}.\n$$\n\nA corollary for ${\\delta}=1$ of Theorem \\ref{thm:LinearWaring} is:\n\\begin{cor}  \\label{cor}\nSuppose a NC homogeneous polynomial $p(x) = \\sum_\\alpha {P_\\alpha x^\\alpha}$, where $P_\\alpha = P_{\\alpha_1, \\alpha_2, ..., \\alpha_d}$ $\\in \\mathbb{C} $,  satisfies $P_\\alpha = P_{\\tilde{\\alpha}}$ for any index sets $\\alpha, \\tilde{\\alpha}$ such that $\\mathbbm{1}_j^\\alpha = \\mathbbm{1}_j^{\\tilde{\\alpha}}$  for all $1 \\le j \\le g$. Then $p$ has an NC complex coefficient Waring decomposition \\ with linear powers.\nMoreover, for such a NC homogeneous polynomial,\nthe number of terms needed is\n$$\n\\left \\lceil \\frac{\\binom{g+d-1}{d}} g  \\right \\rceil,\n$$\nexcept in the cases\n\\begin{itemize}\n     \\item $d = 2$, where $g$ terms are needed\n     \\item $(d,g) = (3,5),(4,3),(4,4),(4,5)$ where $\\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil + 1$ terms are needed.\n \\end{itemize}\n\\end{cor}\n\n\\begin{proof}\nThis corollary is a combination of Theorem \\ref{thm:LinearWaring}, our main result in Section \\ref{main_result}, and the well developed solutions for the\nclassical Waring Theorem stated in Theorem \\ref{classical_result}.\n\\end{proof}\n\n\\subsection{NC polynomial evaluation}\n\nLet $p(x)=\\sum_{|\\alpha| \\leq d} P_\\alpha x^\\alpha$ be a noncommutative polynomial. Then for any $n$ and for any $g$-tuple of $n \\times n$ matrices $X=(X_1, \\dots, X_g)$, we define the \\df{evaluation} of $p$ on $X$ by\n\\[\np(X)=\\sum_{|\\alpha| \\leq d} P_\\alpha X^\\alpha\n\\]\nwhere $X^0=I_n$. In the case where $p$ is a homogeneous noncommutative polynomial and has a NC Waring decomposition, the NC Waring decomposition of $p$ may be used to efficiently evaluate $p$ on matrix tuples. This is especially useful in situations where a single $p$ must be evaluated on many different tuples of matrices.\n\nSuppose $p$ has the NC Waring decomposition\n\\[\np(x)=\\sum_{s=1}^t [A_1^s x_1 + A_2^2 x_2 + \\cdots + A_g^s x_g]^d.\n\\]\nThen for any matrix tuple $X$ we may evaluate $p(X)$ using $tg-1$ matrix additions and $t$ matrix exponentiations of degree $d$, where $t \\leq \\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil + 1$ by Corollary \\ref{cor}. We note that powers of a matrix may be efficiently computed either by decomposing the exponent as a sum of powers of two, or by first computing the Jordan form of the matrix.\n\nIn contrast, a naive evaluation of a single degree $d$ NC monomial requires $d-1$ matrix multiplications. As a consequence, the naive approach to evaluating a homogeneous degree $d$ NC polynomial in $g$ variables on a matrix tuple can require up to $g^d (d-1)$ matrix multiplications and $g^d-1$ matrix additions.\n\nA matrix exponentiation of degree $d$ can easily be evaluated with at most $d-1$ matrix multiplications, so to compare the computation complexity of these methods we may compare $g^d$ and $\\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil + 1$. A more informative comparison on the comparison comes from Stirling’s approximation which shows\n\\[\n\\left\\lceil \\frac 1g  {{g+d-1} \\choose d} \\right\\rceil + 1 \\lessapprox \\frac{1}{g} \\left( \\frac{e(g+d)}{d} \\right)^d.\n\\]\nIt follows that the ratio of $\\lceil \\frac 1g  {{g+d-1} \\choose d} \\rceil$ to $g^d$ is approximately bounded above by\n\\[\n\\frac{1}{g} \\left( \\frac{e(g+d)}{gd} \\right)^d,\n\\]\na quantity that rapidly (geometrically) approaches zero as $g$ or $d$ increase, provided $3 \\leq d,g$.\n\nThe authors thank Ignat Domanov for suggesting efficient polynomial evaluation as an application of NC Waring decompositions.\n\n\n\\subsection{Numerical computation of NC Waring decompositions}\n\\ \\ \\ \\\nWe conclude the introduction with an example which computes an NC Waring decomposition by using popular tensor decomposition software. Consider the homogeneous noncommutative polynomial\n\\[\n\\begin{array}{rclcl}\np(x)&=& x_1^3-4 x_2^3 -4 x_3^3 + 5 x_1 x_1 x_2+ 5 x_1 x_2 x_1 + 5 x_2 x_1 x_1 -3 x_1 x_1 x_3-3 x_1 x_3 x_1-3 x_3 x_1 x_1 \\\\\n& & +7 x_2 x_2 x_1 + 7 x_2 x_1 x_2 + 7 x_1 x_2 x_2-11 x_2 x_2 x_3 - 11 x_2 x_3 x_2 - 11 x_3 x_2 x_2 \\\\\n& & + 6 x_3 x_3 x_1 + 6 x_3 x_1 x_3 +6 x_1 x_3 x_3 - 6 x_3 x_3 x_2 -6 x_3 x_2 x_3-6 x_2 x_3 x_3 \\\\\n& &  + x_1 x_2 x_3 +x_1 x_3 x_2 + x_2 x_1 x_3 +x_2 x_3 x_1+ x_3 x_1 x_2+x_3 x_2 x_1.\n\\end{array}\n\\]\nWe associate $p(x)$ to the symmetric tensor $T$ defined by its frontal slices\n\\[\nT(:,:,1)=\\begin{pmatrix}\n1 & 5 & -3 \\\\\n5 & 7 & 1 \\\\\n-3 & 1 & 6\n\\end{pmatrix}\n\\quad \\quad\n\\mathrm{and}\n\\quad \\quad\nT(:,:,2)=\\begin{pmatrix}\n5 & 7 & 1 \\\\\n7 & -4 & -11 \\\\\n1 & -11 & -6\n\\end{pmatrix}\n\\]\nand\n\\[\nT(:,:,3)=\\begin{pmatrix}\n-3 & 1 & 6 \\\\\n1 & -11 & -6 \\\\\n6 & -6 & -4\n\\end{pmatrix},\n\\]\nwhere $T(:,:,i)$ is the standard Matlab index notation.\n\nUsing Tensorlab \\cite{VDSBL16}  we compute that $T$ is a rank $4$ tensor and has symmetric tensor decomposition\n\\[\nT= v_1 \\otimes v_1 \\otimes v_1 +v_2 \\otimes v_2 \\otimes v_2 +v_3 \\otimes v_3 \\otimes v_3 + v_4 \\otimes v_4 \\otimes v_4\n\\]\nwhere\n\\[\nv_1=\\begin{pmatrix}\n3.839 \\\\\n-1.591 \\\\\n2.593\n\\end{pmatrix}\n\\quad \\quad\nv_2=(-1)^{1/3} \\begin{pmatrix}\n-1.577 \\\\\n-1.697 \\\\\n0.902\n\\end{pmatrix}\n\\]\nand\n\\[\nv_3=\\begin{pmatrix}\n0.821 \\\\\n-2.121 \\\\\n-1.793\n\\end{pmatrix}\n\\quad \\quad\nv_4=\\begin{pmatrix}\n-3.917 \\\\\n1.673 \\\\\n-2.462\n\\end{pmatrix}.\n\\]\nIt follows that $p$ has the rank 4 NC Waring decomposition\n\\[\n\\begin{array}{rclcl}\np(x)&=& (3.839 x_1 - 1.591 x_2 +  2.593 x_3)^3 \\\\\n& & + ((-1)^{1/3} (-1.577 x_1 - 1.697 x_2 + .902 x_3))^3 \\\\\n& & + (.821 x_1 - 2.121 x_2 - 1.793 x_3)^3 \\\\\n& & + (-3.917 x_1 + 1.673 x_2 -2.462 x_3)^3.\n\\end{array}\n\\]\nThis is easy to numerically verify using NCAlgebra \\cite{OHMS}.\n\nA naive evaluation of $p$ on a matrix tuple using the original definition of $p$ requires $54$ matrix multiplications. In contrast, evaluating $p$ on a matrix tuple using its NC Waring decomposition only requires $8$ matrix multiplications.\n\n\\subsection{Guide to readers}\n\\ \\ \\ \\\nIn Section \\ref{sec:linear} we show that the NC Waring problem reduces to the classical Waring problem. The section begins by introducing a compatibility condition which is necessary for a NC homogeneous polynomial $p$ to have a Waring decomposition. The main result of this section is Theorem \\ref{thm:LinearWaring} which shows that a NC homogeneous polynomial $p$ has a $t$-term Waring decomposition if and only if it satisfies our compatibility condition and its commutative collapse has a $t$-term Waring decomposition.\n\nSection \\ref{sec:general} considers the general NC Waring problem. Similar to the $\\delta=1$ case, we begin by introducing a general $\\delta$-compatibility condition which is necessary for the existence of a $(\\delta,d)$-NC Waring decomposition. The main result of the section is Proposition \\ref{prop:GenNCWaringMain} which shows that, under the $\\delta$-compatibility condition, the general NC Waring problem is equivalent to a commutative Waring problem for a polynomial with an increased number of variables. We end with Section \\ref{sec:NeedExtraVars} which illustrates that an increase in our number of variables is necessary to reduce the general NC Waring decomposition \\ to a commutative Waring decomposition .\n\n\n\n\n\n\\section{The  noncommutative Waring problem} \\label{sec:linear}\n\\ \\ \\ \\\nIn this section we will present our main results on the linear noncommutative Waring problem.\n\n\\subsection{Commutative collapse}\n\\ \\ \\ \\\nOur results are associated with commutative problems through a correspondence we now describe.\n\n\nFor a NC polynomial $p$, the associated \\df{commutative collapse}, $p^c$, is the commutative polynomial obtained by considering the variables of $p$ to be commutative.\nOur notation for commutative collapse for an NC monomial $x^\\alpha = x_{\\alpha_1} x_{\\alpha_2} \\dots x_{\\alpha_d}$ is $X^\\alpha  = X_{\\alpha_1} X_{\\alpha_2} \\dots X_{\\alpha_d}$. For example, when $\\alpha = (1, 2, 1, 2 )$, $x^\\alpha = x_1x_2x_1x_2$ collapses to $X^\\alpha = X_1^2X_2^2$.\n\n\nWe impose an equivalence relation $\\sim_c$ \\index{$\\sim_c$} on NC monomials by saying that $x^\\alpha$ and $x^{\\tilde{\\alpha}}$ are \\df{commutative equivalent} if they have the same commutative collapse:\n$$\nx^\\alpha \\sim_c  x^{\\tilde{\\alpha}} \\quad   \\text{iff} \\quad\nX^\\alpha = X^{\\tilde{\\alpha}}.\n$$\nMoreover, two index tuples $\\alpha$ and $\\tilde{\\alpha}$ are \\df{commutative equivalent}, denoted $\\alpha \\sim_c  {\\tilde{\\alpha}}$,\niff $x^\\alpha \\sim_c  x^{\\tilde{\\alpha}}$.\nNote that\n$$\\alpha \\sim_c  {\\tilde{\\alpha}}\\qquad \\mbox{ iff} \\qquad  \\bmone^\\alpha_i = \\mathbbm{1}^{\\widetilde \\alpha}_i\n\\ \\ for \\ i = 1, \\dots, g.$$\n\n\\subsection{Main results on the NC Waring decomposition}\n\\ \\ \\ \\\nOur results contain two parts. First we state a compatibility condition necessary for the existence of a Waring decomposition \\S \\ref{sssec:Compatibility}. Second, if the compatibility condition holds,\nwe reduce the NC Waring problem to the classical commutative\nWaring problem \\S \\ref{main_result}.\n\n\\subsubsection{The Compatibility Condition}\n\\label{sssec:Compatibility}\n\\ \\ \\ \\\nAs we next  see the following condition is necessary for existence\nof a NC Waring decomposition.\n\n\\begin{defi}\nA noncommutative homogeneous degree $d$ polynomial $p(x)$,\n$$\np(x) = \\sum_{|\\alpha|=d} {P_\\alpha x^\\alpha}\\qquad P_\\alpha := P_{\\alpha_1, \\alpha_2, ..., \\alpha_d} \\in \\mathbb{R} \\ or \\ \\mathbb{C},\n$$\nsatisfies the \\df{compatibility condition} means\n\\begin{equation}\n\\label{eq:Paequiv}\nP_\\alpha = P_{\\tilde{\\alpha}} \\qquad for \\ all \\\n\\alpha \\sim_c \\tilde{\\alpha}.\n\\end{equation}\nSometimes we say {$p$ is compatible}.\n\\qed \\end{defi}\n\nA noncommutative homogeneous polynomial $p$ of degree $d$ which satisfies the compatibility condition can be thought of as a ``symmetric\\footnote{The term symmetric typically refers to a noncommutative polynomial $p$ which is equal to its transpose $p^T$, which is different than the notion discussed here.} noncommutative homogeneous polynomial of degree $d$\" in the sense that $p$ is invariant under the following action of the symmetric group. Given tuple a tuple $\\alpha=(\\alpha_1,\\alpha_2, \\dots, \\alpha_d)$ of length $d$ and a permutation $\\pi \\in \\mathcal{S}_d$ define\n\\[\n\\pi (\\alpha)=(\\alpha_{\\pi(1)}, \\alpha_{\\pi(2)}, \\dots, \\alpha_{\\pi(d)}) \\quad \\mathrm{that \\ is} \\quad \\pi (x^\\alpha)=x^{\\pi(\\alpha)}.\n\\]\nIt is then straight forward to check that $x^\\alpha \\sim_c  x^{\\tilde{\\alpha}}$ and $\\alpha \\sim_c \\tilde{\\alpha}$ if and only if there is a permutation $\\pi \\in \\mathcal{S}_d$ such that $\\pi(\\alpha)=\\tilde{\\alpha}$.\n\nWe extend the action of $\\mathcal{S}_d$ to noncommutative homogeneous polynomials of degree $d$ by\n\\[\n\\pi (p(x))=  \\sum_{|\\alpha|=d} {P_\\alpha x^{\\pi(\\alpha)}}.\n\\]\nThen $p$ meets the compatibility condition if and only if\n\\[\n\\pi(p(x))=p(x)\n\\]\nfor all permutations $\\pi \\in \\mathcal{S}_d$.\n\nThe following lemma shows that the compatibility condition is necessary for existence of an NC Waring decomposition.\n\n\n\\begin{lem} \\label{NecessaryCond}\nIf a NC homogeneous polynomial of degree $d$\nhas a  $t$-term NC Waring decomposition,\nthen the compatibility condition \\eqref{eq:Paequiv} holds. Moreover, if $p$ meets the compatibility condition, then $p$ has a $t$-term NC Waring decomposition over the complex numbers (resp. real numbers) if and only if\n\\begin{equation} \\label {Peq}\nP_\\alpha = \\sum_{s=1}^t \\prod_{j=1}^g \\left(  A_j^s \\right)^{\\mathbbm{1}_j^\\alpha}\n\\end{equation}\nhas a solution $A_j^s \\in \\mathbbm{C} ( \\textrm{resp. } A_j^s \\in \\mathbbm{R} ).$\n\\end{lem}\n\n\\begin {proof}\n$p$ has a $t$-term Waring decomposition \\ if and only if\n\\begin{equation*}\n\\label{eq:pcore}\n\\sum_{|\\alpha|=d} P_\\alpha x^\\alpha   = \\sum_{s=1}^t [L_s(x)]^d\n =  \\sum_{s=1}^t   \\sum_{|\\alpha|=d} \\left( \\prod_{i=1}^d A_{\\alpha_i}^s \\right) x^\\alpha\n= \\sum_{|\\alpha|=d} \\left( \\sum_{s=1}^t \\prod_{i=1}^d A_{\\alpha_i}^s \\right) x^\\alpha.\n\\end{equation*}\nComparing the coefficients of $x^\\alpha$ on both sides, we get\n\\begin{equation}\n\\label{eq:pwkeyG}\nP_\\alpha  = \\sum_{s=1}^t \\prod_{i=1}^d A_{\\alpha_i}^s\n= \\sum_{s=1}^t \\prod_{j=1}^g \\left(  A_j^s \\right)^{\\mathbbm{1}_j^\\alpha}\n\\end{equation}\nThis also implies $P_\\alpha = P_{\\tilde{\\alpha}}$ if $\\mathbbm{1}_j^\\alpha\n= \\mathbbm{1}_j^{\\tilde{\\alpha}}$ for all $1 \\le j \\le g $.\n\\end {proof}\n\n\\begin{exa}\n\\rm A NC homogeneous polynomial $p(x) = \\sum_\\alpha {P_\\alpha x^\\alpha}$ has the complex (resp. real) $2$-term Waring decomposition \\\n$$\np(x) = (a x_1 + c x_2)^3 + (b x_1 + d x_2)^3\n$$\nif and only if\n\\begin{equation}\n\\begin{aligned}\nP_{1,1,1}  &\n=   a^3  + b^3\n\\\\\nP_{1,1,2}  &= a^2 c + b^2 d = \\frac 1 6 ((a+c)^3 +(b+d)^3- (a-c)^3-(b-d)^3)- \\frac 1 3 P_{2,2,2}\n\\\\\nP_{1,2,2}  &=   a c^2 + b d^2 = \\frac 1 6 ((a+c)^3 +(b+d)^3+ (a-c)^3+(b-d)^3) - \\frac 1 3 P_{1,1,1}\n\\\\\nP_{2,2,2}  &= c^3  + d^3\n\\end{aligned}\n\\end{equation}\nhas a solution $a,b,c,d \\in \\mathbbm{C}$ (resp. $\\mathbbm{R}$)\n\\qed\n\\end{exa}\n\n\n\n\n\\subsubsection{Reduction of NC Waring to Classical Waring}\n\\label{main_result}\n\\ \\ \\ \\\nThere has been extensive work on the Waring problem in the classical commutative case and what we see in this section is that the NC Waring problem reduces to the commutative one.\n\n\n\n\\begin{lem}\\label{lem:count_eta}\nFor an index tuple $\\alpha$, denote $\\eta[\\alpha]$ as the number of $\\tilde{\\alpha}$'s that satisfy $\\bmone^\\alpha_j = \\bmone^{\\talpha}_j$  for all $1 \\le j \\le g$. Then\n$$\n\\eta[\\alpha] = \\frac {d !} {\\prod_{j = 1}^{g} (\\mathbbm{1}_j^\\alpha) !}\n$$\n\\end{lem}\n\n\\begin{proof}\nThe problem is equivalent to calculating how many d-tuples can be formed by elements from\n$\n\\alpha = ( \\alpha_1, \\alpha_2, \\dots, \\alpha_d )\n$, which is equivalent to\n$$\n\\eta[\\alpha] = \\frac {\\text{\\# of permutations of }d\\text{ items}} {\\text{\\# of permutations of repetitions}} =\\frac {d !} {\\prod_{j = 1}^{g} (\\mathbbm{1}_j^\\alpha) !}.\n$$\n\\end{proof}\n\n\\begin{thm} \\label{thm:LinearWaring}\nSuppose $p$ is an NC homogeneous polynomial which satisfies the compatibility conditions \\eqref{eq:Paequiv}.\nThen the commutative collapse $p^c$ has the Waring decomposition \\\n\\begin{equation} \\label{eq:comm_WR}\np^c(X)=\\sum_{s=1}^t [ A^s_1X_1 + A^s_2X_2 + \\cdots + A^s_g X_g]^d\n\\end{equation}\n(with $X_i$ being commuting variables) if and only if $p$ has the\nnc Waring decomposition \\\n\\begin{equation} \\label{eq:nc_WR}\np(x)=\\sum_{s=1}^t [ A^s_1x_1 + A^s_2x_2 + \\cdots +A^s_g x_g]^d .\n\\end{equation}\nNote that the number of terms is the same and the\nreal coefficients\n(resp. complex coefficients) $A_j^s$  are the same.\n\\end{thm}\n\n\\def{\\mathcal R}{{\\mathcal R}}\n\n\n\\begin{proof}\nThe proof begins by laying out the algebraic connection\nbetween $p$ and $p^c$. Let ${\\mathcal R}$ denote a set consisting of one  representative from each $\\sim_c$ equivalence class. Then from \\eqref{eq:Paequiv}, the NC polynomial\n$\np(x) = \\sum_{|\\alpha| = d}  P_\\alpha x^\\alpha\n$\nhas\ncommutative collapse satisfying\n$$\np^c(X)\n =\n \\sum_{\\alpha \\in {\\mathcal R}  }\n\\quad  \\sum_{ \\tilde{\\alpha} \\sim_c  \\alpha } P_{\\tilde{\\alpha}} X^{\\alpha}\n =\n \\sum_{\\alpha \\in {\\mathcal R} }\n P_\\alpha^c \\  X^{\\alpha},\n$$\nwhere\n$\nP_\\alpha^c =\n \\sum_{ \\tilde{\\alpha} \\sim_c  \\alpha } P_{\\tilde{\\alpha}}\n$.\n\nThus if $p$ satisfies the compatibility condition\n \\eqref{eq:Paequiv}, then\n\\begin{equation}\n \\label{ncToc}\nP_\\alpha^c =  \\eta[\\alpha ]    P_{\\widetilde \\alpha} \\qquad for \\ {\\alpha} \\in {\\mathcal R}\n\\ \\ and \\ \\\n{\\alpha} \\sim_c {\\widetilde \\alpha}.\n\\end{equation}\nTherefore, $p^c$ is the commutative collapse of a\ncompatible NC homogeneous degree $d$ polynomial $p$\n iff $P_\\alpha^c =  \\eta[\\alpha ]    P_\\alpha$ for all index tuples $\\alpha$ of length $d$.\n\nNow we proceed to prove our theorem. Assume $p$ has the NC Waring decomposition \\ \\eqref{eq:nc_WR}, we shall obtain a reversible formula for the Waring decomposition \\ of $p^c$.\nBy equation \\eqref{ncToc} and Lemma \\ref{NecessaryCond}, the commutative collapse $p^c$ is\n\\begin{equation} \\label{eq:collapse}\np^c(X) =\n \\sum_{\\alpha \\in {\\mathcal R}, \\  | \\alpha | = d  }\\eta[\\alpha ]    P_\\alpha X^\\alpha  = \\sum_{|\\alpha|=d} P_\\alpha X^\\alpha  = \\sum_{|\\alpha|=d} \\sum_{s=1}^t \\prod_{j=1}^g \\left( A_j^s \\right)^{\\mathbbm{1}_j^\\alpha} X^\\alpha\n\\end{equation}\nThus\n\\begin{equation} \\label{eq:collapse2}\np^c(X)\n  = \\sum_{s=1}^t \\sum_{|\\alpha|=d} \\prod_{i=1}^d A^s_{\\alpha_i} X^\\alpha\n  = \\sum_{s=1}^t [ A^s_1X_1 + A^s_2X_2 + ... A^s_g X_g]^d\n\\end{equation}\n\n\n\n\nOn the other hand, suppose $p$'s commutative collapse, $p^c$, has the commutative Waring decomposition \\ \\eqref{eq:comm_WR}, then\nthe calculations in \\eqref{eq:collapse} and \\eqref{eq:collapse2} can be reversed. By comparing coefficients, this is equivalent to\n$$\nP^c_\\alpha = \\eta[\\alpha ]\n \\sum_{s=1}^t \\prod_{j=1}^g \\left( A_j^s \\right)^{\\mathbbm{1}_j^\\alpha}\n$$\nfor all $\\alpha \\in {\\mathcal R}$.\nTherefore by \\eqref{ncToc},\n$p$ satisfies\n $$\nP_\\alpha =\\sum_{s=1}^t \\prod_{j=1}^g \\left( A_j^s \\right)^{\\mathbbm{1}_j^\\alpha}\n$$\nfor all index tuples $\\alpha$ of length $d$. Hence by Lemma \\ref{NecessaryCond}, $p$ has the Waring decomposition \\ \\eqref{eq:nc_WR}.\nThus under the compatibility condition \\eqref{eq:Paequiv}, the NC polynomial $p$ has a Waring decomposition \\ iff its commutative collapse $p^c$ has the same Waring decomposition.\n\\end{proof}\n\n\n\n\\section {The general noncommutative Waring problem} \\label{sec:general}\n\\ \\ \\ \\\n We now consider a more general situation of which the problem in the preceding section is the base case. As you will see, the bookkeeping and notation is formidable, so it is very helpful to have done a simpler case. In the previous section our focus was to determine if a degree $d$ noncommutative \\  homogeneous polynomial can be expressed as sums of powers of linear terms. Now we examine when a degree ${\\delta} d$ noncommutative \\  homogeneous polynomial can be expressed as sums of powers of homogeneous degree ${\\delta}$ terms.\n\n\n\\subsection{Problem formulation and notation}\n\\ \\ \\ \\\nLet \\index{$T^g_\\delta$} $T^g_\\delta$ be the set of all possible $\\delta$-tuples whose elements are integers between $1$ and $g$, i.e.,\n$$\nT^g_\\delta = \\{ (\\alpha^1, \\alpha^2, \\dots, \\alpha^\\delta) \\mid 1 \\le \\alpha^i \\le g \\}.\n$$\nAdditionally, define \\index{$( T^g_\\delta )^d $} $( T^g_\\delta )^d $ by\n$$\n( T^g_\\delta )^d = \\{ (\\alpha_1, \\alpha_2, \\dots, \\alpha_d) \\mid \\alpha_i \\in T^g_\\delta \\}.\n$$\nThat is, $( T^g_\\delta )^d$ is the set of $d$-tuples of $\\delta$ tuples of indices.\nFor any $\\alpha = (\\alpha_1, \\dots, \\alpha_d) \\in (T^g_\\delta)^d$, where $\\alpha_i = (\\alpha_i^1, \\dots, \\alpha_i^{\\delta}) \\in T^g_\\delta, $\nwe can write\n\\[\nx^{\\alpha} = x^{\\alpha_1}x^{\\alpha_2} \\dots x^{\\alpha_d}.\n\\]\nIt is natural to identify $ x^\\alpha$ with\n\\[\nx_{\\alpha_1^1} x_{\\alpha_1^2} \\dots x_{\\alpha_1^\\delta} x_{\\alpha_2^1}\n\\dots x_{\\alpha_d^1}\n\\dots x_{\\alpha_d^\\delta}.\n\\]\n\nRecall our notation for a degree $\\delta$ homogeneous polynomial is\n$$\nH(x) = \\sum_{\\beta \\in T^g_\\delta} A_{\\beta} x^\\beta,\n$$\nwhere $A_{\\beta} = A_{(\\beta^1, \\beta^2, \\dots, \\beta^\\delta)} \\in \\mathbb{C}$.\n\n\\begin{rmk}\n\\rm For any $ \\alpha = (\\alpha_1, \\alpha_2, \\dots, \\alpha_d) \\in (T^g_\\delta)^d$,  we can identify\n$$\\alpha = ((\\alpha_1^1, \\alpha_1^2, \\dots, \\alpha_1^\\delta), \\dots, (\\alpha_d^1, \\alpha_d^2, \\dots, \\alpha_d^\\delta))$$\nwith\n$$\n (\\alpha_1^1, \\alpha_1^2, \\dots, \\alpha_1^\\delta, \\dots, \\alpha_d^\\delta) \\in T^g_{\\delta d}.$$ On the other hand, for any element of $T^g_{\\delta d}$, we can reverse this identification and form groups of size $\\delta$ to get a $d$-tuple of $\\delta$-tuples. Hence $(T^g_\\delta)^d$ and $T^g_{\\delta d}$ are isomorphic\nand we let $\\tau$ \\index{$\\tau$} denote the isomorphism\n $$\\tau : T^g_{\\delta d}  \\to (T^g_\\delta)^d$$\n which accomplishes this grouping.\n\\qed\n\\end{rmk}\n\n\\bigskip\n\n\\noindent\n{\\bf The} \\df{General NC Waring Problem:}\n\n {\\bf\n\\ \\ \\ \\ Given a NC homogeneous degree $\\delta d$ polynomial p, does it have a t-term $d^{th}$ power real NC Waring (resp. complex NC Waring)\ndecomposition of  degree $\\delta$. That is, can $p(x)$ be written as\n\\begin{equation}\n\\label{eq:deldWaring}\np(x) = \\sum_{s = 1}^t (H_s(x))^d = \\sum_{s = 1}^t \\left ( \\sum_{\\beta \\in T^g_\\delta} A^s_{\\beta} x^{\\beta} \\right )^d?\n\\end{equation}}\n\n\\bigskip\n\nWe call this problem the $({\\delta},d)$-NC Waring problem and say a a decomposition of the form \\eqref{eq:deldWaring} is a \\df{$t$-term $({\\delta},d)$-NC Waring decomposition}. Similarly for a commutative polynomial $p^c$, we say a decomposition of the form \\eqref{eq:deldWaring} (with $x^\\beta$ replaced by $X^\\beta$) is a \\df{$t$-term $({\\delta},d)$-Waring decomposition}. Note that the problem treated in Section \\ref{sec:linear} is exactly the $(1,d)$-NC Waring problem.\n\nAn obvious fact is, if $p$ is a degree $\\delta d$ NC homogeneous polynomial and $p$ has a $t$-term $(\\delta,d)$-NC Waring decomposition ,\nthen its commutative collapse $p^c$ has a $t$-term $(\\delta,d)$-Waring decomposition .\n\n\\subsubsection{Tuple indicator functions}\n\\ \\ \\ \\\nWe now extend the notion of indicator function to tuples of ${\\delta}$-tuples.\nFor two $\\delta-$tuples $\\beta,\\gamma \\in T^g_{\\delta}$, denote\n\\index{$\\mathbbm{1}_\\beta^{\\gamma}$ }\n$$\n\\mathbbm{1}_\\beta^{\\gamma} =\n\\begin{cases}\n1 & \\textrm{if }  \\gamma = \\beta \\\\\n0 & \\textrm{otherwise},\n\\end{cases}.\n$$\nThen for an index tuple $\\mu \\in (T^g_\\delta)^d$, the number of times a particular $\\delta-$tuple $\\beta \\in T^g_\\delta$ appears in $\\mu$ is\n$$\n\\mathbbm{1}_\\beta^\\mu := \\sum_{k=1}^d \\mathbbm{1}_\\beta^{\\mu_k}.\n$$\nFurthermore, denote\n\n\n\\begin{equation}\\label{eq:1abi}\n\\mathbbm{1}_i^\\mu\n:= \\sum_{\\beta \\in T^g_\\delta, i \\in \\beta}\n\\mathbbm{1}^\\mu_\\beta\n= \\sum_{\\beta \\in T^g_\\delta} \\mathbbm{1}^\\mu_\\beta \\bmone^\\beta_i\n\\end{equation}\nas the number of integers $i$ appearing in all the $\\delta$-tuples in $\\alpha$.\n\n\n\n\\subsection{Main results on the general Waring decomposition}\n\\ \\ \\ \\\nSimilar to Section 2, we first state a compatibility\ncondition which is necessary for the existence of a generalized NC Waring decomposition . We then prove that, if this condition holds, then we can reduce the generalized NC Waring problem to a commutative one at the price of an increase our number of variables.\n\n\n\n\\subsubsection{The Compatibility Condition}\n\\ \\ \\ \\\nThe generalized version of the $\\delta = 1$ compatibility condition is defined as follows:\n\\begin{defi}\nA noncommutative homogeneous polynomial of degree $\\delta d$ in g variables of the form\n\n\\begin{equation} \\label{eq:ncPoly}\np(x) = \\sum_{\\alpha \\in {T^g_{\\delta d} }} {P_\\alpha x^\\alpha} \\qquad P_\\alpha\n \\in \\mathbb{R} \\ or \\ \\mathbb{C}\n\\end{equation}\nsatisfies the \\df {$\\delta$-compatibility condition} means\n\\begin{equation} \\label{eq:GeneralPaEquiv}\nP_{\\alpha} = P_{\\tilde{\\alpha}}\n\\end{equation}\nfor all index sets, $\\alpha$, $\\tilde{\\alpha} \\in {T^g_{\\delta d} }$\nsuch that $\\bmone^{\\tau(\\alpha)}_\\beta = \\bmone^{\\tau(\\tilde \\alpha) }_\\beta \\text{ for all }\n\\beta \\in T^g_\\delta $.\nConsistent with this,\nwe define the \\df{$\\delta$-equivalence relation}, denoted $\\sim_\\delta$ \\index{$\\sim_\\delta$}, on ${T^g_{\\delta d} }$ by\n\n$$\n\\alpha \\sim_{\\delta} \\tilde{\\alpha}\n\\quad\n\\text{iff}\n\\quad\n\\bmone^{\\tau(\\alpha)}_\\beta  = \\bmone^{\\tau(\\tilde \\alpha) }_\\beta\n$$\nfor all $\\beta \\in T^g_\\delta$.\n\\qed \\end{defi}\n\n\n\n\n\n\\begin{rmk}\n\\label{rem:tupequiv}\n\\rm Here are a few bookkeeping properties of $\\delta$-equivalences.\n\n\\begin{enumerate}\n\n\\item\n\\label{it:1EquivIsCommEquiv}\nWe have $\\alpha \\sim_1 \\tilde{\\alpha}$ if and only if $\\alpha \\sim_c \\tilde{\\alpha}$.\n\n\\item\n\\label{it:monEquivCompare}\nLet $\\delta_1,\\delta_2 \\in \\mathbb{N}$ and let $\\alpha,\\tilde{\\alpha} \\in T_{\\delta_2 d}^g$. If $\\delta_2$ divides $\\delta_1$, then $\\alpha \\sim_{\\delta_1} \\tilde{\\alpha}$ implies $\\alpha \\sim_{\\delta_2} \\tilde{\\alpha}$. In the case where $\\delta_2=1$ this follows from equation \\eqref{eq:1abi}. The general case is similar.\n\n\\item\n\\label{it:polyEquivCompare}\nLet $\\delta_1, \\delta_2, d \\in \\mathbb{N}$ and let $p$ be a degree $\\delta_1 d$ NC homogeneous polynomial. If $\\delta_2$ divides $\\delta_1$ and $p$ satisfies the $\\delta_2$-compatibility condition then $p$ satisfies the $\\delta_1$-compatibility condition.\n\\end{enumerate}\nItems \\eqref{it:monEquivCompare} and \\eqref{it:polyEquivCompare} highlight that, as $\\delta$ grows, it becomes increasingly difficult for fixed monomials $\\alpha$ and $\\tilde{\\alpha}$ of degree divisible by $\\delta$ to be $\\delta$-equivalent. As an immediate consequence, as $\\delta$ grows, it become more likely that a fixed NC homogeneous polynomial $p$ of degree divisible by $\\delta$ satisfies the $\\delta$-compatibility condition. In the extreme case, monomials $\\alpha$ and $\\tilde{\\alpha}$ of degree $\\delta$ are $\\delta$-equivalent if and only if $\\alpha=\\tilde{\\alpha}$. As a result, every degree $\\delta$ NC homogeneous polynomial satisfies the $\\delta$-compatibility condition. \\qed\n\n\\end{rmk}\n\n\\begin{exa}\n\\rm Let\n\\[\n\\alpha=x_1 x_2 x_2 x_1 \\quad \\quad \\quad \\quad \\tilde{\\alpha}=x_2 x_1 x_1 x_2.\n\\]\nThen\n\\[\n\\alpha \\sim_1 \\tilde{\\alpha} \\quad \\mathrm{and} \\quad  \\alpha \\sim_2 \\tilde{\\alpha} \\quad  \\mathrm{however} \\quad  \\alpha \\not\\sim_4 \\tilde{\\alpha}.\n\\]\nNow let $p$ be the degree four homogeneous NC polynomial\n\\[\np(x)=\\alpha+\\tilde{\\alpha}=x_1 x_2 x_2 x_1+x_2 x_1 x_1 x_2.\n\\]\nThen $p$ satisfies the $2$-compatibility condition and the $4$-compatibility condition. However, $p$ does not satisfy the $1$-compatibility condition, since the coefficient of $x_1 x_1 x_2 x_2$ in $p$ is $0$ but the coefficient of $x_1 x_2 x_2 x_1$ is $1$ and\n\\[\nx_1 x_1 x_2 x_2 \\sim_1 x_1 x_2 x_2 x_1. \\qed\n\\]\n\\end{exa}\n\n\n\n The following lemma shows that the $\\delta$-compatibility condition is necessary for the general NC Waring problem.\n\n\\begin{lem}\n\\label{lem:condgen}\nSuppose a NC homogeneous polynomial $p$ of degree\n$\\delta d$ in $g$ variables has a $t$-term $({\\delta},d)$-NC Waring decomposition ,\n then $p$  satisfies the $\\delta$-compatibility condition,\n $P_{\\alpha} = P_{\\widetilde \\alpha}$  if ${\\alpha} \\sim_\\delta {\\widetilde \\alpha}$.\n Here $p$ has  coefficients  $P_{\\alpha}$.\n\nMoreover, the $({\\delta},d)$-NC\n Waring problem  has a\nsolution over the complex numbers (resp. real numbers) if and only if the equation\n\\begin{equation} \\label {eq:pcoef}\nP_\\alpha = \\sum_{s=1}^t \\prod_{\\beta \\in T^g_\\delta}\n\\left (A^s_\\beta \\right)^{\\bmone^{\\tau(\\alpha)}_\\beta}\n\\qquad {\\alpha} \\in T_{{\\delta} d}^g\n\\end{equation}\nhas a solution $A_\\beta^s \\in \\mathbbm{C}$ (resp. $A_\\beta^s \\in \\mathbbm{R}$).\n\\end{lem}\n\n\\begin{proof}\n$p$ has a $t$-term $(\\delta,d)$-NC Waring decomposition \\,\niff $\\exists$ $\\delta^{th}$ degree  homogeneous polynomials,\n$H_1, H_2, \\dots, H_t$ satisfying\n\\begin{align}\n\\sum_{\\alpha \\in {T^g_{\\delta d} }} P_\\alpha x^\\alpha\n& = \\sum_{s=1}^t [H_s(x)]^d\n  = \\sum_{s=1}^t \\left [\\sum_{\\beta \\in T^g_\\delta}\n  A_{\\beta}^s x^{\\beta} \\right ]^d\\\\\n  %\n& = \\sum_{s=1}^t   \\sum_{\\alpha \\in {T^g_{\\delta d} } }\n \\left( \\prod_{ \\substack { 1 \\le j \\le d}}\n A_{\\tau(\\alpha)_j}^s x^{\\alpha_j}\\right) \\\\\n & = \\sum_{\\alpha \\in {T^g_{\\delta d} }}\n \\left( \\sum_{s=1}^t \\prod_{ \\substack { 1 \\le j \\le d}}\n A_{\\tau(\\alpha)_j}^s \\right) x^\\alpha.\n \\label{eq:pmain}\n\\end{align}\n\\newline\nComparing coefficients we see, equivalent to the $(\\delta, d)$-NC Waring\ndecomposition is:\n$$\nP_\\alpha  = \\sum_{s=1}^t \\prod_{ \\substack {\n1 \\le j \\le d}} A_{\\tau(\\alpha)_j}^s\n= \\sum_{s=1}^t \\prod_{\\substack{{j_1, \\dots,j_\\delta }\\\\\n1 \\le j_k \\le g}}^g\n\\left(  A_{(j_1, \\dots,j_\\delta)}^s \\right)^{\\mathbbm{1}_{(j_1, \\dots,j_\\delta)}^{\\tau(\\alpha)}}\n= \\sum_{s=1}^t \\prod_{\\beta \\in T^g_\\delta} \\left (A^s_\\beta \\right)^{\\bmone^{\\tau(\\alpha)}_\\beta},\n$$\nyielding \\eqref{eq:pcoef}.\n\nAs a consequence\n$\nP_\\alpha = P_{\\tilde{\\alpha}}\n$\nfor any\n$\\alpha$\nsatisfying\n$\\bmone^{\\tau(\\alpha)}_\\beta = \\bmone^{\\tau(\\tilde \\alpha) }_\\beta$\nfor every\n$\n\\beta \\in T^g_\\delta\n$, yielding the first assertion of the theorem.\n\\end{proof}\n\n\n\\begin{exa}\n\\rm\nLet\n$$\n    p(x) = (x_1x_2 + x_1^2)(x_2x_1 + x_1^2)\n      = x_1x_2^2x_1 + x_1x_2x_1^2 + x_1^2x_2x_1 + x_1^4.\n$$\nThen $p$ is an example where there is no $({\\delta},d)=(2,2)$-NC\nWaring decomposition; indeed\nthe $2$-compatibility condition is violated because\n$\nP_{(1,1,1,2)} = 0 \\neq 1 = P_{(1,2,1,1)}\n$.\nHowever, its commutative collapse does have the Waring decomposition :\n\\begin{equation*}\n    p^c(X)=X_1^2X_2^2 + 2 X_1^3 X_2 + X_1^4\n          = (X_1X_2 + X_1^2)^2.  \\qquad \\qed\n\\end{equation*}\n\n\\end{exa}\n\n\n\\subsection{Reduction to classical Waring in more variables}\n\\label{sec:GenWaringReduction}\n\\ \\ \\ \\\n\nTo solve the general $(\\delta,d)$-noncommutative \\  Waring  problem we reduce to the  $\\delta=1$ case solved by Theorem \\ref{thm:LinearWaring}.\nThis reduction is accomplished by identifying a monomial $x^\\beta$\nwith a new variable $z_\\beta$. Namely, fix $\\delta$ and define the map $\\phi$ on monomials of the form $x^\\beta$ for $\\beta \\in T_\\delta^g$ by\n$$\n\\phi (x^\\beta) := z_\\beta \\qquad for \\ each \\ \\beta \\in T_\\delta^g\n$$\nwhere the $z_\\beta$ are noncommutative indeterminates indexed by elements of $T_\\delta^g$.\n\nWe extend or definition of $\\phi$ to a noncommutative homogeneous polynomial\n\\[\np(x)=\\sum_{\\mu \\in (T_\\delta^g)d} P_\\mu x^{\\mu_1} x^{\\mu_2} \\cdots x^{\\mu_d}\n\\]\n of degree $\\delta d$\n by\n\\begin{equation}\n\\label{eq:phiOnPolys}\n\\phi(p(x))=\\sum_{\\mu \\in (T_\\delta^g)d} P_\\mu \\phi(x^{\\mu_1}) \\phi(x^{\\mu_2}) \\cdots \\phi(x^{\\mu_d})=\\sum_{\\mu \\in (T_\\delta^g)d} P_\\mu z_{\\mu_1} z_{\\mu_2}  \\cdots z_{\\mu_d}.\n\\end{equation}\n\n\\begin{lem}\n\\label{lem:phiAlgIso}\nThe map $\\phi$ as defined in equation \\eqref{eq:phiOnPolys} defines an algebra isomorphism on the algebra of noncommutative homogeneous polynomials of degree divisible by $\\delta$ in the noncommutative indeterminate $x=(x_1,x_2, \\dots, x_g)$ which maps to the algebra of noncommutative homogeneous polynomials in the noncommutative indeterminates $\\{z_\\beta\\}_{\\beta \\in T_\\delta^g}$.\n\\end{lem}\n\\begin{proof}\nThis is straight forward from the definition of $\\phi$ on a noncommutative homogeneous polynomial of degree $d\\delta$.\n\\end{proof}\n\nWe now give our main result for the $(\\delta,d)$-NC Waring problem.\n\n\\begin{prop}\n\\label{prop:GenNCWaringMain}\nLet $p$ be a noncommutative homogeneous polynomial of degree $\\delta d$ in the indeterminate $x=(x_1, \\dots, x_g)$, and let $\\phi$ be as defined in equation \\eqref{eq:phiOnPolys}. Then we have the following.\n\\begin{enumerate}\n\\item\n\\label{it:GenMain1}\n$p(x)$ has a $t$-term  $({\\delta},d)$-noncommutative \\  Waring decomposition if and only if\n$\\phi(p(x))$ has a $t$-term $(1,d)$-noncommutative Waring decomposition.\n\n\\item\n\\label{it:GenMain2}\n$p(x)$ satisfies the ${\\delta}$-compatibility condition if and only if $\\phi(p(x))$ satisfies the  $1$-compatibility condition.\n\n\\item\n\\label{it:GenMain3}\n$p(x)$ has a $t$-term  $({\\delta},d)$-noncommutative \\  Waring decomposition if and only if $p(x)$ satisfies the ${\\delta}$-compatibility condition  and the commutative collapse of $\\phi(p(x))$ has a $t$-term $(1,d)$-Waring decomposition.\n\\end{enumerate}\n\\end{prop}\n\\begin{proof}\nTo prove item \\eqref{it:GenMain1}, assume $p(x)$ has a $t$-term $(\\delta,d)$-noncommutative Waring decomposition\n\\[\np(x)= \\sum_{s=1}^t  [\\sum_{\\beta \\in T_{\\delta}^g } A_\\beta \\; x^\\beta  ]^d.\n\\]\nBy Lemma \\ref{lem:phiAlgIso}, $\\phi$ is an algebra isomorphism so\n\\[\n\\phi(p(x))=\\phi\\left(\\sum_{s=1}^t  [\\sum_{\\beta \\in T_{\\delta}^g } A_\\beta \\; x^\\beta  ]^d \\right) = \\sum_{s=1}^t  [\\sum_{\\beta \\in T_{\\delta}^g } A_\\beta \\; \\phi(x^\\beta)  ]^d = \\sum_{s=1}^t  [\\sum_{\\beta \\in T_{\\delta}^g } A_\\beta \\; z_\\beta  ]^d.\n\\]\nThis shows $\\phi(p(x))$ has a $t$-term $(1,d)$ noncommutative Waring decomposition. The reverse direction is follows the same reasoning using $\\phi^{-1}$ instead of $\\phi$.\n\nTo prove item \\eqref{it:GenMain2} let\n\\[\np(x)=\\sum_{\\mu \\in (T_\\delta^g)d} P_\\mu x^{\\mu_1} x^{\\mu_2} \\cdots x^{mu_d}.\n\\]\nThen\n\\[\n\\phi(p(x))=\\sum_{\\mu \\in (T_\\delta^g)d} P_\\mu z_{\\mu_1} z_{\\mu_2}  \\cdots z_{\\mu_d}.\n\\]\nObserve\n\\[\n(\\mu_1, \\dots, \\mu_d) \\sim_1 (\\tilde \\mu_1, \\dots \\tilde \\mu_d)\n\\]\nwhere the $\\mu_j$ are viewed as elements of the index set $T_\\delta^g$ if and only if\n\\[\n(\\mu_1, \\dots, \\mu_d) \\sim_{\\delta} (\\tilde \\mu_1, \\dots \\tilde \\mu_d)\n\\]\nwhere the $\\mu_j$ are viewed as as $\\delta$ tuples of elements of $T_{\\delta}^g$. It follows that\n\\[\nP_{(\\mu_1, \\dots, \\mu_d)}=P_{(\\tilde \\mu_1, \\dots \\tilde \\mu_d)} \\quad \\quad \\mathrm{for\\ all\\ } (\\mu_1, \\dots, \\mu_d) \\sim_1 (\\tilde \\mu_1, \\dots \\tilde \\mu_d).\n\\]\nwhere the $\\mu_j$ are viewed as elements of the index set $T_\\delta^g$ if and only if\n\\[\nP_{(\\mu_1, \\dots, \\mu_d)}=P_{(\\tilde \\mu_1, \\dots \\tilde \\mu_d)} \\quad \\quad \\mathrm{for\\ all\\ } (\\mu_1, \\dots, \\mu_d) \\sim_{\\delta} (\\tilde \\mu_1, \\dots \\tilde \\mu_d).\n\\]\nwhere the $\\mu_j$ are viewed as as $\\delta$ tuples of elements of $T_{\\delta}^g$.\n\n\nItem \\eqref{it:GenMain3} is an immediate consequence of items \\eqref{it:GenMain1} and \\eqref{it:GenMain2} with Theorem \\ref{thm:LinearWaring}, our main result for $(1,d)$-NC  Waring decompositions.\n\\end{proof}\n\n\\subsection{Additional variables are necessary for the reduction}\n\\label{sec:NeedExtraVars}\n\\ \\ \\ \\\nIt is tempting to try to solve the general $({\\delta},d)$-NC Waring problem by reducing to the commutative case without introducing additional variables. This section will show that this is not possible.\n\nOne may hope that the following are true:\n\n\\begin{enumerate}\n\\item\n\\label{it:comm_WR_general}\n\\textit{If $p$ is a degree $\\delta d$ NC homogeneous polynomial, which satisfies the $\\delta$-compatibility condition \\eqref{eq:GeneralPaEquiv},\nthen its commutative collapse $p^c$ has the Waring decomposition \\\n\\begin{equation} \\label{comm_WR_general}\np^c(X) = \\sum_{s = 1}^t \\left ( \\sum_{\\beta \\in T^g_\\delta} A^s_{\\beta} X^{\\beta} \\right )^d\n\\end{equation}\n(with $X_i$ being commuting variables) if and only if $p$ has the\nNC Waring decomposition \\\n\\begin{equation} \\label{nc_WR_general}\np(x)=\\sum_{s=1}^t \\left ( \\sum_{\\beta \\in T^g_\\delta} A^s_{\\beta} x^{\\beta} \\right )^d.\n\\end{equation}\n}\n\n\\item\n\\label{it:nc_WR_general}\n\\textit{The commutative collapse $p^c$ of $p$\nhas a $t$-term  $({\\delta},d)$-NC Waring decomposition\niff\nthe commutative collapse\n$\\phi(p)^c$ of $\\phi(p)$ has a $t$-term\n$(1,d)$-NC Waring decomposition.}\n\n\\end{enumerate}\n\nThe following polynomial gives a counter example to both items. Let\n\\[\np(x)=x_1^4+x_1 x_2 x_2 x_1 + x_2 x_1 x_1 x_2 +x_2^4\n\\]\nand let $\\delta=d=2$. Then $p$ satisfies the $2$-compatibility condition. We will show that the commutative collapse of $p$ has a two term $(2,2)$-Waring decomposition \\  but that $p$ does not have a two term $(2,2)$-NC  Waring decomposition.\n\nIt is straight forward to check\n\\[\np^c(X)=X_1^4+2X_1^2 X_2^2 +X_2^4=(X_1^2+X_2^2)^2.\n\\]\nItem \\eqref{it:comm_WR_general} would imply that\n\\[\np(x)=(x_1^2+x_2^2)^2=x_1^4+x_1^2 x_2^2+x_2^2 x_1^2 +x_2^4 \\neq x_1^4+x_1 x_2 x_2 x_1 + x_2 x_1 x_1 x_2 +x_2^4 = p(x)\n\\]\nwhich is contradiction. This shows that item \\eqref{it:comm_WR_general} cannot be correct.\n\nIn fact, $p$ does not have a two term $(2,2)$-NC  Waring decomposition. To check this set\n\\[\nz_{(1,1)}=x_1 x_1 \\quad \\quad z_{(1,2)}= x_1 x_2 \\quad \\quad z_{(2,1)}=x_2 x_1 \\quad \\quad z_{(2,2)}=x_2 x_2.\n\\]\nThen $\\phi(p) (z)=z_{(1,1)}^2+z_{(1,2)} z_{(2,1)} +z_{(2,1)} z_{(1,2)} +z_{(2,2)}^2$ satisfies the $1$-compatibility condition but\n\\[\n(\\phi(p))^c (Z)= Z_{(1,1)}^2+ 2 Z_{(1,2)} Z_{(2,1)} + Z_{(2,2)} ^2\n\\]\ndoes not have a two term $(1,2)$-Waring decomposition.\\footnote{This is a straight forward calculation.}\nIt follows from Proposition \\ref{prop:GenNCWaringMain} \\eqref{it:GenMain3} that $p$ does not have a two term $(2,2)$-NC Waring decomposition.\n\n", "meta": {"timestamp": "2019-03-15T01:15:45", "yymm": "1903", "arxiv_id": "1903.05910", "language": "en", "url": "https://arxiv.org/abs/1903.05910", "abstract": "This paper analyses a Waring type decomposition of a noncommuting (NC) polynomial $p$ with respect to the goal of evaluating $p$ efficiently on tuples of matrices. Such a decomposition can reduce the number of matrix multiplications needed to evaluate a noncommutative polynomial and is valuable when a single polynomial must be evaluated on many matrix tuples.In pursuit of this goal we examine a noncommutative analog of the classical Waring problem and various related decompositions. For example, we consider a \"Waring decomposition\" in which each product of linear terms is actually a power of a single linear NC polynomial or more generally a power of a homogeneous NC polynomial. We describe how NC polynomials compare to commutative ones with regard to these decompositions, describe a method for computing the NC decompositions and compare the effect of various decompositions on the speed of evaluation of generic NC polynomials.", "subjects": "Functional Analysis (math.FA)", "title": "Efficient evaluation of noncommutative polynomials using tensor and noncommutative Waring decompositions", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9838471628097781, "lm_q2_score": 0.85776809953619, "lm_q1q2_score": 0.843912711077416}}
{"url": "https://arxiv.org/abs/1608.01666", "title": "A central limit theorem for a new statistic on permutations", "abstract": "This paper does three things: It proves a central limit theorem for novel permutation statistics (for example, the number of descents plus the number of descents in the inverse). It provides a clear illustration of a new approach to proving central limit theorems more generally. It gives us an opportunity to acknowledge the work of our teacher and friend B. V. Rao.", "text": "\\section{Introduction}\nLet $S_n$ be the group of all $n!$ permutations of $\\{1,\\ldots, n\\}$. A variety of statistics $T(\\pi)$ are used to enable tasks such as tests of randomness of a time series, comparison of voter profiles when candidates are ranked, non-parametric statistical tests and evaluation of search engine rankings. A basic feature of a permutation is a local `up-down' pattern. Let the number of descents be defined~as \n\\[\nD(\\pi) := |\\{i: 1\\le i\\le n-1,\\, \\pi(i+1)<\\pi(i)\\}|\\,.\n\\]\nFor example, when $n=10$, the permutation $\\pi = (\\underline{7} \\; 1\\; \\underline{5}\\; 3\\; \\underline{10}\\; \\underline{8}\\; \\underline{6}\\; 2\\; 4\\; 9)$ has $D(\\pi)= 5$. The enumerative theory of permutations by descents has been intensively studied since Euler. An overview is in Section \\ref{overview} below. In seeking to make a metric on permutations using descents we were led to study\n\\begin{equation}\\label{tpi}\nT(\\pi) := D(\\pi)+D(\\pi^{-1})\\,.\n\\end{equation}\nFor a statistician or a probabilist it is natural to ask ``Pick $\\pi\\in S_n$ uniformly; what is the distribution of $T(\\pi)$?'' While a host of limit theorems are known for $D(\\pi)$, we found $T(\\pi)$ challenging. A main result of this paper establishes a central limit theorem. \n\\begin{thm}\\label{clt}\nFor $\\pi$ chosen uniformly from the symmetric group $S_n$, and $T(\\pi)$ defined by \\eqref{tpi}, for $n\\ge 2$\n\\[\n\\mathbb{E}(T(\\pi)) = n-1\\,,\\; \\; \\mathrm{Var}(T(\\pi)) = \\frac{n+7}{6}-\\frac{1}{n}\\,,\n\\]\nand, normalized by its mean and variance, $T(\\pi)$ has a limiting standard normal distribution.\n\\end{thm}\nThe proof of Theorem \\ref{clt} uses a method of proving central limit theorems for complicated functions of independent random variables due to Chatterjee~\\cite{cha08}. This seems to be a useful extension of Stein's approach. Indeed, we were unable to prove Theorem~\\ref{clt} by standard variations of Stein's method such as dependency graphs, exchangeable pairs or size-biased couplings. Theorem \\ref{clt} is a special case of the following more general result. Call a statistic $F$ on $S_n$ ``local of degree $k$'' if $F$ can be expressed as\n\\[\nF(\\pi) = \\sum_{i=0}^{n-k} f_i(\\pi)\\,,\n\\] \nwhere the quantity $f_i(\\pi)$ depends only on the relative ordering of $\\pi(i+1),\\ldots, \\pi(i+k)$. For example, the number of descents is local of degree $2$, and the number of peaks is local of degree $3$. We will refer to $f_0,\\ldots, f_{n-k}$ as the ``local components'' of $F$. \n\\begin{thm}\\label{genclt}\nSuppose that $F$ and $G$ are local statistics of degree $k$ on $S_n$, as defined above. Suppose further that the absolute values of the local components of $F$ and $G$ are uniformly bounded by $1$. Let $\\pi$ be be chosen uniformly from $S_n$ and let \n\\[\nW := F(\\pi) + G(\\pi^{-1})\\,.\n\\]\nLet $s^2 := \\mathrm{Var}(W)$. Then the Wasserstein distance between $(W-\\mathbb{E}(W))/s$ and the standard normal distribution is bounded by $C(n^{1/2}s^{-2} + n s^{-3}) k^3$, where $C$ is a universal constant. \n\\end{thm}\nAfter the first draft of this paper was posted on arXiv, it was brought to our notice that the joint normality of  $D(\\pi)$, $D(\\pi^{-1})$ was proved in Vatutin~\\cite{vatutin96} in 1996 via a technical tour de force with generating functions. The asymptotic normality in Theorem \\ref{clt} follows as a corollary of Vatutin's theorem. Theorem \\ref{genclt} is a new contribution of this paper.\n\nIn outline, Section \\ref{metrics} describes metrics on permutations and our motivation for the study of $T(\\pi)$. Section \\ref{overview} reviews the probability and combinatorics of $D(\\pi)$ and $T(\\pi)$. Section \\ref{method} describes Chatterjee's central limit theorem. Section \\ref{proof} proves Theorems \\ref{clt} and the proof of Theorem \\ref{genclt} is in Section \\ref{proof2}. Section \\ref{future} outlines some other problems where the present approach should work.\n\\vskip.2in\n\\noindent{\\bf Acknowledgments.} This work derives from conversations with Ron Graham. Walter Stromquist provided the neat formula for the variance of $T(\\pi)$. Jason Fulman provided some useful references. Vladimir Vatutin brought the important reference \\cite{vatutin96} to our notice. Finally, B.~V.~Rao has inspired both of us by the clarity, elegance and joy that he brings to mathematics.\n\n\n\n\n\n\\section{Metrics on permutations}\\label{metrics}\nA variety of metrics are in widespread use in statistics, machine learning, probability, computer science and the social sciences. They are used in conjunction with statistical tests, evaluation of election results (when voters rank order a list of candidates), and for combing results of search engine rankings. The book by Marden \\cite{marden} gives a comprehensive account of various approaches to statistics on permutations. The book by Critchlow \\cite{critchlow} extends the use of metrics on permutations to partially ranked data (top $k$ out of $n$) and other quotient spaces of the symmetric group. The Ph.D.~thesis of Eric Sibony (available on the web) has a comprehensive review of machine learning methods for studying partially ranked data. Finally, the book by Diaconis \\cite[Chapter 6]{diaconis} contains many metrics on groups and an extensive list of applications. \n\nSome widely used metrics are:\n\\begin{itemize}\n\\item Spearman's footrule: $d_s(\\pi, \\sigma) = \\sum_{i=1}^n |\\pi(i)-\\sigma(i)|$.\n\\item Spearman's rho: $d_\\rho^2(\\pi, \\sigma) = \\sum_{i=1}^n (\\pi(i)-\\sigma(i))^2$. \n\\item Kendall's tau: $d_\\tau(\\pi,\\sigma) = $ minimum number of adjacent transpositions to bring $\\sigma$ to $\\pi$. \n\\item Cayley: $d_C(\\pi, \\sigma) = $ minimum number of transpositions to bring $\\sigma$ to $\\pi$. \n\\item Hamming: $d_H(\\pi, \\sigma) = |\\{i: \\pi(i)\\ne \\sigma(i)\\}|$. \n\\item Ulam: $d_U(\\pi,\\sigma) = n - \\text{length of longest increasing subsequence in } \\pi\\sigma^{-1}$.\n\\end{itemize}\nAll of these have known means, variances and limit laws \\cite{diaconis}. Some of this is quite deep mathematically. For example, the limit law for Ulam's metric is the Tracy--Widom distribution of random matrix theory.\n\nIn addition to the metric properties, metrics can be classified by their invariance properties; for example, right invariance ($d(\\pi, \\sigma) = d(\\pi\\eta, \\sigma\\eta)$), left invariance ($d(\\pi, \\sigma) = d(\\eta \\pi, \\eta \\sigma)$), two-sided invariance ($d(\\pi, \\sigma) = d(\\eta_1\\pi\\eta_2, \\eta_1\\sigma\\eta_2)$), and conjugacy invariance ($d(\\pi, \\sigma) = d(\\eta^{-1}\\pi\\eta, \\eta^{-1}\\sigma\\eta)$). Common sense requires right invariance; if $\\pi$ and $\\sigma$ are the rankings of a class on the midterm and on the final, we would not want $d(\\pi,\\sigma)$ to depend on the class being listed by last name or identity number. All of the metrics above are right invariant. Only the Cayley and Hamming metrics are bi-invariant.\n\\vskip.1in\n\\noindent{\\underline{A metric from descent structure?}}\n\nIt is natural to try and make a metric from descents. Let us call this $d_D(\\pi, \\sigma)$. By right invariance only the distance $d_D(\\text{id}, \\sigma)$ must be defined (and then $d_D(\\pi, \\sigma) = d_D(\\text{id}, \\sigma\\pi^{-1})$). A zeroth try is $d^0_D(\\text{id}, \\sigma) = D(\\sigma)$; at least $d_D^0(\\text{id}, \\text{id}) = 0$. However, symmetry requires $d_D^0(\\text{id}, \\sigma) = d^0_D(\\sigma, \\text{id}) = d_D^0(\\text{id}, \\sigma^{-1})$ and $D(\\sigma)\\ne D(\\sigma^{-1})$ for many $\\sigma$ (for example, when $n=4$, $\\sigma = (2\\; 4\\; 1\\; 3)$, $\\sigma^{-1}= (3\\; 1\\; 4\\; 2)$, $D(\\sigma) = 1$, $D(\\sigma^{-1})=2$). In small samples $D(\\sigma)=D(\\sigma^{-1})$ occurs fairly often. However, in Section \\ref{proof} we will prove a bivariate central limit theorem for $D(\\sigma)$, $D(\\sigma^{-1})$, which suggests that the chance of $D(\\sigma)=D(\\sigma^{-1})$ is asymptotic to $Cn^{-1/2}$ when $n$ is large. \n\nA next try is $d_D^1(\\text{id}, \\sigma) = D(\\sigma)+D(\\sigma^{-1})$. Then $d_D^1(\\pi, \\sigma) = D(\\sigma \\pi^{-1}) + D(\\pi\\sigma^{-1}) = d_D^1(\\sigma, \\pi)$. Alas, Ron Graham showed us simple examples where this definition fails to satisfy the triangle inequality! Take $\\pi = (3\\;4\\;1\\;2\\;5)$, $\\sigma =(1\\;4\\;5\\;2\\;3)$. A simple check shows that\n\\[\n2+2 = d_D^1(\\pi, \\text{id}) + d_D^1(\\text{id}, \\sigma) < d_D^1(\\pi, \\sigma) = 6\\,.\n\\]\nThe next idea does work. Form a graph with vertices the $n!$ permutations and an edge from $\\pi$ to $\\sigma$ with weight $D(\\pi\\sigma^{-1}) + D(\\sigma \\pi^{-1})$.  Define $d_D^2(\\pi, \\sigma)$ as the minimum sum of the weights of paths from $\\pi$ to $\\sigma$. Experiments show that {\\it usually} the minimum path is the edge from $\\pi$ to $\\sigma$. But the example above shows this is not always the case. We believe that for almost all pairs the graph distance equals the edge weight.\n\nThe statistic $T(\\pi) = D(\\pi)+D(\\pi^{-1})$ arose from these considerations. Of course, $T$ does not have to give rise to a metric to be a useful measure of disarray. The Kullback--Leibler `divergence' is a case in point. \n\n\n\\section{Combinatorics and probability for descents}\\label{overview}\nThe study of descents starts with Euler. In studying power series which allow closed form evaluation, Euler showed that \n\\begin{equation}\\label{euler}\n\\sum_{k=0}^\\infty k^n t^k = \\frac{A_n(t)}{(1-t)^{n+1}}\\,,\n\\end{equation}\nwhere the Eulerian polynomial is\n\\[\nA_n(t) = \\sum_{\\pi\\in S_n} t^{D(\\pi)+1} = \\sum_{i=1}^n A_{n,i} t^i\\,,\n\\]\nwith $A_{n,i} = |\\{\\pi\\in S_n: D(\\pi)=i-1\\}|$, the Eulerian numbers. Thus,\n\\[\n\\sum_{k=0}^\\infty t^k = \\frac{1}{1-t}\\,,\\;\\;\\sum_{k=0}^\\infty kt^k = \\frac{t}{(1-t)^2}\\,, \\;\\;\\sum_{k=0}^\\infty k^2 t^k = \\frac{t+t^2}{(1-t)^3}\\, , \\;\\;\\ldots\n\\]\nFulman \\cite{fulman99} connects \\eqref{euler} and other descent identities to `higher math'.\n\nThe Eulerian numbers and polynomials satisfy a host of identities and have neat generating functions. We recommend Carlitz~\\cite{carlitz}, Petersen~\\cite{petersen13},  Stanley~\\cite[p.\\ 6]{stanley77}, Graham--Knuth--Patashnik~\\cite[Chapter 6]{grahametal}, the book by Petersen~\\cite{petersen15} and the references in Sloane~\\cite[Seq A008292]{sloane} for basics with pointers to a large literature. \n\nThe probability theory of descents is similarly well studied. The central limit theorem reads:\n\\begin{thm}\\label{descent}\nFor $\\pi$ chosen uniformly in $S_n$, \n\\[\n\\mathbb{E}(D(\\pi)) = \\frac{n-1}{2}\\,,\\;\\;\\mathrm{Var}(D(\\pi)) = \\frac{n+1}{12}\\,,\n\\]\nand, normalized by its mean and variance, $D(\\pi)$ has a limiting standard normal distribution.\n\\end{thm}\n\\noindent{\\it \\underline{Remark.}} We point here to six different approaches to the proof of Theorem \\ref{descent}. Each comes with an error term (for the Kolmogorov distance) of order $n^{-1/2}$. The first proof uses $m$-dependence: For $1\\le i\\le n-1$, let \n\\[\nX_i(\\pi) = \n\\begin{cases}\n1&\\text{ if } \\pi(i+1)<\\pi(i),\\\\\n0 &\\text{ else.}\n\\end{cases}\n\\]\nIt is easy to see that $X_1,\\ldots, X_{n-1}$ are $2$-dependent. The central limit theorem follows. See Chen and Shao \\cite{chenshao} for a version with error terms. A second proof uses a geometrical interpretation due to Stanley~\\cite{stanley77}. Let $U_1,\\ldots, U_n$ be independent uniform random variables on $[0,1]$. Stanley shows that for all $n$ and $0\\le j\\le n-1$, \n\\[\n\\mathbb{P}(D(\\pi)=j) = \\mathbb{P}(j < U_1+\\cdots +U_n < j+1)\\,.\n\\]\nFrom here the classical central limit theorem for sums of i.i.d.~random variables gives the claimed result. A third proof due to Harper~\\cite{harper} uses the surprising fact that the generating function $A_n(t)$ has all real zeros. Now, general theorems for such generating functions show that $D(\\pi)$ has the same distribution as the sum of $n$ independent Bernoulli random variables with success probabilities determined by the zeros. Pitman \\cite{pitman} surveys this topic and shows\n\\[\n\\sup_{-\\infty<x<\\infty}\\left|\\mathbb{P}\\left(\\frac{D(\\pi)-\\frac{n-1}{2}}{\\sqrt{\\frac{n+1}{12}}} \\le x\\right) - \\Phi(x)\\right|\\le \\sqrt{\\frac{12}{n}}\\,,\n\\]\nwhere $\\Phi$ is the standard normal cumulative distribution function.\n\nA fourth approach due to Fulman \\cite{fulman04} uses a clever version of Stein's method of exchangeable pairs (see also Conger \\cite{conger}). Each of the papers cited above gives pointers to yet other proofs. A related probabilistic development is in Borodin, Fulman and Diaconis~\\cite{borodinetal}. They show that the descent process is a determinantal point process and hence the many general theorems for determinantal point processes apply.\n\nThe fifth approach is due to Bender~\\cite{bender}, who uses generating functions to prove the CLT. Lastly, David and Barton~\\cite{davidbarton} give a proof using the method of moments.\n\nWe make two points regarding the above paragraphs. First, all of the proofs depend on some sort of combinatorial magic trick. Second, we were unable to get any of these techniques to work for~$T(\\pi)$. \n\nThere have been some applications of descents and related local structures (for example, peaks) in statistics. This is nicely surveyed in Warren and Seneta~\\cite{warrenseneta} and Stigler \\cite{stigler}.\n\\vskip.1in\n\\noindent{\\underline{Joint distribution of $D(\\pi)$, $D(\\pi^{-1})$}}\n\nThere have been a number of papers that study the joint distribution of $D(\\pi)$, $D(\\pi^{-1})$ (indeed, along with other statistics). We mention Rawlings~\\cite{rawlings} and Garsia and Gessel~\\cite{garsiagessel}. As shown below, these papers derive generating functions in a sufficiently arcane form that we have not seen any way of deriving the information we need from them. \n\nTwo other papers seem more useful. The first by Kyle Petersen \\cite{petersen13} treats only $D(\\pi)$, $D(\\pi^{-1})$ and is very accessible. The second, by Carlitz, Rosellet and Scoville~\\cite{carlitzetal}, gives useful recurrences via `manipulatorics'. \n\nFirst, let\n\\[\nA_{n,r,s} = |\\{\\pi\\in S_n: D(\\pi)=r-1,\\, D(\\pi^{-1}) = s-1\\}| \\,.\n\\]\nA table of $A_{8,r,s}$ from Peterson \\cite[p.~12]{petersen13} shows a roughly elliptic shape and suggests that the limiting distribution of $D(\\pi)$, $D(\\pi^{-1})$ might be a bivariate normal distribution with vanishing correlations. Theorem \\ref{clt2} from Section \\ref{proof} proves normality with limiting correlation zero. \n\nCarlitz et al.~\\cite{carlitzetal} delineate which $r$, $s$ are possible:\n\\[\nA_{n,r,s}=0 \\iff r\\ge \\frac{s-1}{s} n + 1\\,.\n\\]\nLet \n\\[\nA_n(u,v) =  \\sum_{\\pi\\in S_n} u^{D(\\pi)+1} v^{D(\\pi^{-1})+1} = \\sum_{i,j=1}^{n} A_{n,i,j} u^i v^j\\,.\n\\]\nPetersen \\cite[Theorem 2]{petersen13} gives the following formula for the generating function defined above:\n\\[\n\\frac{A_n(u,v)}{(1-u)^{n+1}(1-v)^{n+1}} = \\sum_{k,l\\ge 0} {kl+n-1\\choose n} u^k v^l\\,.\n\\]\nFrom this, he derives a recurrence which may be useful for deriving moments:\n\\begin{align*}\nnA_n(u,v) &= (n^2uv + (n-1)(1-u)(1-v))A_{n-1}(u,v)\\\\\n&\\qquad + nuv(1-u)\\fpar{}{u}A_{n-1}(u,v) + nuv(1-v) \\fpar{}{v} A_{n-1}(u,v)\\\\\n&\\qquad + uv(1-u)(1-v) \\mpar{}{u}{v} A_{n-1}(u,v)\\,.\n\\end{align*}\nFrom the above identity, he derives a complicated recurrence for $A_{n,i,j}$.  We are interested in $D(\\pi)+D(\\pi^{-1})$, so the relevant generating function is\n\\[\nA_n(u,u) = (1-u)^{2n+2} \\sum_{k,l\\ge 0} {kl+n-1\\choose n} u^{k+l}\\,.\n\\] \nFinally, Carlitz et al.~\\cite{carlitzetal} give\n\\[\n\\sum_{n=0}^\\infty\\sum_{i=0}^\\infty \\sum_{j=0}^\\infty A_{n,i,j} z^ix^jy^n(1-x)^{-(n+1)} (1-z)^{-(n+1)} = \\sum_{k=0}^\\infty \\frac{z^k}{1-x(1-y)^{-k}}\\,.\n\\]\nComplicated and seemingly intractable as they are, the generating functions displayed above are in fact amenable to analysis. In a remarkable piece of work from twenty years ago, Vatutin \\cite{vatutin96} was able to use these generating functions to prove a class of multivariate central limit theorems for functions of $\\pi$ and $\\pi^{-1}$. The results were generalized to other settings in Vatutin~\\cite{vatutin94} and Vatutin and Mikhailov~\\cite{vm96}.\n\n\n\n\\section{The method of interaction graphs}\\label{method}\nOne motivation for the present paper is to call attention to a new approach to proving central limit theorems for non-linear functions of independent random variables. Since most random variables can be so presented, the method has broad scope. The method is presented in \\cite{cha08} in abstract form and used to solve a spatial statistics problem of Bickel; this involved `locally dependent' summands where the notion of local itself is determined from the data. A very different application is given in~\\cite{chasound}. We hope that the applications in Sections \\ref{proof} and \\ref{future} will show the utility of this new approach.\n\n\nThe technique of defining `local neighborhoods' using the data was named `the method of interaction graphs' in \\cite{cha08}. The method can be described very briefly as follows.  \nLet $\\mathcal{X}$ be a set endowed with a sigma algebra and let $f:\\mathcal{X}^n \\rightarrow \\mathbb{R}$ be a measurable map, where $n\\ge 1$ is a given positive integer. Suppose that $G$ is a map that associates to every $x\\in \\mathcal{X}^n$ a simple graph $G(x)$ on $[n] := \\{1,\\ldots,n\\}$. Such a map will be called a graphical rule on $\\mathcal{X}^n$. We will say that a graphical rule $G$ is symmetric if for any permutation $\\pi$ of $[n]$ and any $(x_1,\\ldots,x_n)\\in \\mathcal{X}^n$, the set of edges in $G(x_{\\pi(1)}, \\ldots, x_{\\pi(n)})$ is exactly\n\\[\n\\{\\{\\pi(i),\\pi(j)\\}: \\{i,j\\} \\text{ is an edge of } G(x_1,\\ldots,x_n)\\}.\n\\]\nFor $m \\ge n$, a symmetric graphical rule $G'$ on $\\mathcal{X}^m$ will be called an extension  of $G$ if for any $(x_1,\\ldots,x_m) \\in \\mathcal{X}^m$, $G(x_1,\\ldots,x_n)$ is a subgraph of $G'(x_1,\\ldots,x_m)$.\n\nNow take any $x,x'\\in \\mathcal{X}^n$. For each $i\\in [n]$, let $x^i$ be the vector obtained by replacing $x_i$ with $x'_i$ in the vector $x$. For any two distinct elements $i$ and $j$ of~$[n]$, let $x^{ij}$ be the vector obtained by replacing $x_i$ with $x'_i$ and $x_j$ with~$x_j'$. We will say that the coordinates $i$ and $j$ are non-interacting for the triple $(f,x,x')$ if\n\\[\nf(x)-f(x^j) = f(x^i) - f(x^{ij}).\n\\]\nWe will say that a graphical rule $G$ is an interaction rule for a function $f$ if for any choice of $x,x'$ and $i,j$, the event that $\\{i,j\\}$ is not an edge in the graphs $G(x)$, $G(x^i)$, $G(x^j)$, and $G(x^{ij})$ implies that $i$ and $j$ are non-interacting vertices for the triple $(f,x,x')$.  The following theorem implies central limit behavior if one can construct an interaction graph that has, with high probability, small maximum degree.\n\\begin{thm}[\\cite{cha08}]\\label{main}\nLet $f:\\mathcal{X}^n \\rightarrow \\mathbb{R}$ be a measurable map that admits a symmetric interaction rule $G$. \nLet $X_1,X_2,\\ldots $ be a sequence of i.i.d.\\ $\\mathcal{X}$-valued random variables and let $X= (X_1,\\ldots,X_n)$. Let $W := f(X)$ and $\\sigma^2 := \\mathrm{Var}(W)$. Let $X^\\prime = (X^\\prime_1,\\ldots,X^\\prime_n)$ be an independent copy of $X$. For each $j$, define\n\\[\n\\Delta_j f(X) = W - f(X_1,\\ldots,X_{j-1}, X^\\prime_j, X_{j+1}, \\ldots,X_n),\n\\]\nand let $M = \\max_j|\\Delta_jf(X)|$. Let $G'$ be an extension of $G$ on $\\mathcal{X}^{n+4}$, and put\n\\[\n\\delta := 1+ \\textup{degree of the vertex $1$ in } G'(X_1,\\ldots,X_{n+4}).\n\\]\nThen \n\\[\n\\delta_W \\le \\frac{Cn^{1/2}}{\\sigma^2}\\mathbb{E}(M^8)^{1/4} \\mathbb{E}(\\delta^4)^{1/4} +  \\frac{1}{2\\sigma^3}\\sum_{j=1}^n \\mathbb{E}|\\Delta_j f(X)|^3,\n\\]\nwhere $\\delta_W$ is the Wasserstein distance between $(W-\\mathbb{E}(W))/\\sigma$ and $N(0,1)$, and $C$ is a universal  constant.\n\\end{thm}\n\n\n\\section{Proof of Theorem \\ref{clt}}\\label{proof}\nWe now apply Theorem \\ref{main} to prove Theorem \\ref{clt}. Take $\\mathcal{X} = [0,1]^2$, and let $X_1,X_2,\\ldots$ be independent uniformly distributed points on $\\mathcal{X}$. Let $X= (X_1,\\ldots, X_n)$. Write each $X_i$ as a pair $(U_i,V_i)$. Define the $x$-rank of the point $X_i$ as the rank of $U_i$ among all the $U_j$'s, and the $y$-rank of the point $X_i$ as the rank of $V_i$ among all the $V_j$'s. More accurately, we should say ``$x$-rank of $X_i$ in $X$'' and ``$y$-rank of $X_i$ in $X$''.\n\nLet $X_{(1)},\\ldots,X_{(n)}$ be the $X_i$'s arranged according to their $x$-ranks, and let $X^{(1)},\\ldots,X^{(n)}$ be the $X_i$'s arranged according to their $y$-ranks. Write $X_{(i)} = (U_{(i)}, V_{(i)})$ and $X^{(i)} = (U^{(i)}, V^{(i)})$. Let $\\pi(i)$ be the $y$-rank of $X_{(i)}$. Then clearly $\\pi$ is a uniform random permutation. Let $\\sigma(i)$ be the $x$-rank of $X^{(i)}$. Then $X^{(i)} = X_{(\\sigma(i))}$. Therefore \n\\[\n\\pi(\\sigma(i)) = \\text{the $y$-rank of $X_{(\\sigma(i))}$} = \\text{the $y$-rank of $X^{(i)}$} = i\\,.\n\\]\nThus, $\\sigma = \\pi^{-1}$. Let \n\\begin{align*}\nW = f(X) &:=  \\sum_{i=1}^{n-1} 1_{\\{\\pi(i)> \\pi(i+1)\\}} + \\sum_{i=1}^{n-1} 1_{\\{\\sigma(i)>\\sigma(i+1)\\}}\\\\\n&= \\sum_{i=1}^{n-1} 1_{\\{V_{(i)} > V_{(i+1)}\\}} + \\sum_{i=1}^{n-1}1_{\\{U^{(i)} > U^{(i+1)}\\}}\\,.\n\\end{align*}\nNow suppose that one of the $X_i$'s is replaced by an independent copy $X_i'$. Then $W$ can change by at most $4$. Therefore in the notation of Theorem \\ref{main}, $|\\Delta_j f(X)|\\le 4$ for every $j$ and hence $M\\le 4$.  \n\nFor $x\\in \\mathcal{X}^n$, define a simple graph $G(x)$ on $[n]$ as follows. For any $1\\le i\\ne j\\le n$, let $\\{i,j\\}$ be an edge if and only if the $x$-rank of $x_i$ and the $x$-rank of $x_j$ differ by at most $1$, or the $y$-rank of $x_i$ and the $y$-rank of $x_j$ differ by at most $1$. This construction is clearly invariant under relabeling of indices, and hence $G$ is a symmetric graphical rule. \n\nGiven $x,x'\\in \\mathcal{X}^n$, let $x^i$, $x^j$ and $x^{ij}$ be defined as in the paragraphs preceding the statement of Theorem \\ref{main}. Suppose that $\\{i,j\\}$ is not an edge in the graphs $G(x)$, $G(x^i)$, $G(x^j)$ and $G(x^{ij})$. Since the difference $f(x)-f(x^i)$ is determined only by those $x_k$'s such that $k$ is a neighbor of $i$ in  $G(x)$ or in $G(x^i)$, and the difference $f(x^j)-f(x^{ij})$ is determined only by those $x_k$'s such that $k$ is a neighbor of $i$ in $G(x^j)$ or in $G(x^{ij})$, the above condition implies that $i$ and $j$ are non-interacting for the triple $(f,x,x')$. Thus, $G$ is a symmetric interaction rule for $f$. \n\n\nNext, define a graphical rule $G'$ on $\\mathcal{X}^{n+4}$ as follows. For any $x\\in \\mathcal{X}^{n+4}$ and $1\\le i\\ne j\\le n+4$, let $\\{i,j\\}$ be an edge if and only if the $x$-rank of $x_i$ and the $x$-rank of $x_j$ differ by at most $5$, or if the $y$-rank of $x_i$ and the $y$-rank of $x_j$ differ by at most $5$. Then $G'$ is a symmetric graphical rule. Since the addition of four extra points can alter the difference between the ranks of two points by at most $4$, any edge of $G(x_1,\\ldots,x_n)$ is also an edge of $G'(x_1,\\ldots,x_{n+4})$. Thus, $G'$ is an extension of $G$. \n\nSince the degree of $G'(x)$ is bounded by $10$ for any $x\\in \\mathcal{X}^{n+4}$, the quantity $\\delta$ of Theorem \\ref{main} is bounded by $10$ in this example. The upper bounds on $\\Delta_j f(X)$, $M$ and $\\delta$ obtained above imply that for this $W$,\n\\[\n\\delta_W\\le \\frac{Cn^{1/2}}{\\sigma^2} + \\frac{Cn}{\\sigma^3}\\,,\n\\]\nwhere $\\sigma^2 = \\mathrm{Var}(W)$ and $\\delta_W$ is the Wasserstein distance between $(W-\\mathbb{E}(W))/\\sigma$ and $N(0,1)$.\n\nTo complete the proof of Theorem \\ref{clt}, we derive the given expression for $\\mathrm{Var}(T(\\pi))$. We thank Walter Stromquist for the following elegant calculation. From classically available generating functions, it is straightforward to show that \n\\[\n\\mathbb{E}(D(\\pi)) = \\frac{n-1}{2}\\,,\\;\\;\\mathrm{Var}(D(\\pi)) = \\frac{n+1}{12}\\,,\n\\]\nas given in Theorem \\ref{descent} (see Fulman \\cite{fulman04}). Consider, with obvious notation,\n\\[\n\\mathbb{E}(D(\\pi)D(\\pi^{-1})) = \\mathbb{E}\\biggl(\\sum_{i=1}^{n-1} \\sum_{j=1}^{n-1}D_i(\\pi)D_j(\\pi^{-1})\\biggr)\\,.\n\\]\nFor a given $\\pi$, the $(n-1)^2$ terms in the sum may be broken into three types, depending on the size of the set $\\{\\pi(i), \\pi(i+1)\\} \\cap \\{j,j+1\\}$, which may be $0$, $1$ or $2$. Carefully working out the expected value in each case, it follows that\n\\[\n\\mathbb{E}(D(\\pi)D(\\pi^{-1})) =  \\frac{(n-1)^2}{4} + \\frac{n-1}{2n}\\,.\n\\]\nFrom this,\n\\[\n\\mathrm{Cov}(D(\\pi), D(\\pi^{-1})) = \\frac{n-1}{2n},\\ \\ \\ \\mathrm{Var}(D(\\pi)+D(\\pi^{-1})) = 2\\biggl(\\frac{n+1}{12}\\biggr) + 2\\biggl(\\frac{n-1}{2n}\\biggr) = \\frac{n+7}{6}-\\frac{1}{n}\\,.\n\\]\nThis completes the proof of Theorem \\ref{clt}.\n\\vskip.1in\n\\noindent{\\it \\underline{Remarks.}} The argument used to prove Theorem \\ref{clt} can be adapted to prove the joint limiting normality of $(D(\\pi), D(\\pi^{-1}))$. Fix real $(s,t)$ and consider $T_{s,t}(\\pi)=s D(\\pi)+t D(\\pi^{-1})$. The same symmetric graph $G$ can be used. Because $s$, $t$ are fixed, the limiting normality follows as above. The limiting variance follows from $\\mathrm{Cov}(D(\\pi), D(\\pi^{-1}))$. We state the conclusion:\n\\begin{thm}\\label{clt2}\nFor $\\pi$ chosen uniformly from the symmetric group $S_n$, \n\\[\n\\left(\\frac{D(\\pi)-\\frac{n-1}{2}}{\\sqrt{\\frac{n+1}{12}}},\\, \\frac{D(\\pi^{-1})-\\frac{n-1}{2}}{\\sqrt{\\frac{n+1}{12}}}\\right) \\stackrel{d}{\\longrightarrow} Z_2\\,,\n\\]\nwhere $Z_2$ is bivariate normal with mean $(0,0)$ and identity covariance matrix.\n\\end{thm}\nLike Theorem \\ref{clt}, Theorem \\ref{clt2} can also be deduced as a corollary of the results in \\cite{vatutin96}.\n\n\n\\section{Proof of Theorem \\ref{genclt}}\\label{proof2}\nThe proof goes exactly as the proof of Theorem \\ref{clt}, except for slight differences in the definitions of the interaction graphs $G$ and $G'$. Recall the quantity $k$ from the statement of Theorem \\ref{genclt}. With all the same notation as in the previous section, define the interaction graph $G$ as follows. For any $1\\le i\\ne j\\le n$, let $\\{i,j\\}$ be an edge if and only if the $x$-rank of $x_i$ and the $x$-rank of $x_j$ differ by at most $k-1$, or the $y$-rank of $x_i$ and the $y$-rank of $x_j$ differ by at most $k-1$. It is easy to see that this is indeed an interaction graph for $F(\\pi)+G(\\pi^{-1})$. Define the extended graph $G'$ similarly, with $k-1$ replaced by $k+3$. Everything goes through as before, except that the quantity $M$ of Theorem \\ref{main} is now bounded by $C_1 k$ (since the local components of $F$ and $G$ are bounded by $1$) and the quantity $\\delta$ of Theorem \\ref{main} is now bounded by $C_2 k$, where $C_1$ and $C_2$ are universal constants. The proof is now easily completed by plugging in these bounds in Theorem \\ref{main}.\n\n\\section{Going further}\\label{future}\nThere are two directions where the methods of this paper should allow natural limit theorems to be proved. The first is to develop a similar theory for {\\it global} patterns in subsequences (for example triples $i<j<k$ with $\\pi(i)<\\pi(j)<\\pi(k)$). Of course, inversions, the simplest case ($i<j$ with $\\pi(i)>\\pi(j)$), have a nice probabilistic and combinatorial theory, so perhaps the joint distribution of the number of occurrences of a fixed pattern in $\\pi$ and $\\pi^{-1}$ does as well. \n\nThe second direction of natural generalization is to other finite reflection groups. There is a notion of `descents' --- the number of simple positive roots sent to negative roots. The question of number of number of descents in $x$ plus the number of descents in $x^{-1}$ makes sense and present methods should apply. In specific contexts, for example the hyperoctahedral group of signed permutations, these are concrete interesting questions. \n\nLastly, it would be interesting to see if the dependence of the error bound on $k$ in Theorem \\ref{genclt} can be improved, and to figure out what is the optimal dependence.\n\n\n\n\n\n", "meta": {"timestamp": "2016-10-28T02:01:23", "yymm": "1608", "arxiv_id": "1608.01666", "language": "en", "url": "https://arxiv.org/abs/1608.01666", "abstract": "This paper does three things: It proves a central limit theorem for novel permutation statistics (for example, the number of descents plus the number of descents in the inverse). It provides a clear illustration of a new approach to proving central limit theorems more generally. It gives us an opportunity to acknowledge the work of our teacher and friend B. V. Rao.", "subjects": "Probability (math.PR)", "title": "A central limit theorem for a new statistic on permutations", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9875683465856102, "lm_q2_score": 0.8539127548105611, "lm_q1q2_score": 0.8432972073966294}}
{"url": "https://arxiv.org/abs/1805.01412", "title": "Upper bounds for the regularity of powers of edge ideals of graphs", "abstract": "Let $G$ be a finite simple graph and $I(G)$ denote the corresponding edge ideal. In this paper, we obtain upper bounds for the Castelnuovo-Mumford regularity of $I(G)^q$ in terms of certain combinatorial invariants associated with $G$. We also prove a weaker version of a conjecture by Alilooee, Banerjee, Beyarslan and Hà on an upper bound for the regularity of $I(G)^q$ and we prove the conjectured upper bound for the class of vertex decomposable graphs. Using these results, we explicitly compute the regularity of $I(G)^q$ for several classes of graphs.", "text": "\\section{Introduction}\nLet $I$ be a homogeneous ideal of a polynomial ring \n$R = \\K[x_1,\\ldots,x_n]$ over a field $\\K$ with usual grading.\nIn \\cite{BEL91}, Bertram, Ein and Lazarsfeld  \nhave initiated  the study of the Castelnuovo-Mumford regularity of $I^q$\nas  a function of $q$ by  proving that if $I$ is the defining ideal of a smooth complex \nprojective variety, then $\\reg(I^q)$ is bounded by a linear function of $q$. \nThen, Chandler \\cite{Cha97} and Geramita, Gimigliano and Pitteloud \\cite{GGP95} proved that\nif $\\dim(R/I)\\leq 1$, then $\\reg(I^q) \\leq q.\\reg(I)$ for all $q\\geq 1$. \nHowever, Swanson \\cite{Swa97} proved that there exists $k \\geq 1$ such that\nfor all $q \\geq 1$, $\\reg(I^q)\\leq kq$. Thereafter, Cutkosky, Herzog and Trung, \\cite{CHT}, \nand independently Kodiyalam \\cite{vijay}, proved that for a homogeneous ideal \n$I$ in a polynomial ring,  $\\reg(I^q)$ is a linear \nfunction for $q \\gg 0$ i.e., \nthere exist non negative integers $a$ and $b$ depending on $I$ such that\n$\\reg(I^q)=aq+b \\text{ for all $q \\gg 0$}.$\nWhile the coefficient $a$ is well-understood (\\cite{CHT}, \\cite{vijay}, \\cite{TW}), \nthe free constant $b$ and the stabilization index $q_0=\\min\\{q' \\mid \\reg(I^q)=aq+b, \n\\text{ for all } q \\geq q'\\}$ are quite mysterious.\nTherefore, the attention has been to identify classes for which the linear\npolynomial can be computed or bounded using invariants associated to $I$.\nThere have been some attempts on computing the free constant and stabilization index\nfor several class of ideals. For instance, if $I$  is a equigenerated homogeneous ideal, then \n$b$ is related to the regularity of fibers of certain projection map \n(see for example, \\cite{Romer2001}). If\n$I$ is $(x_1,\\ldots,x_n)$-primary, then $q_0$ can be related to partial\nregularity of the Rees algebra of $I$ (see for example, \\cite{Berle}).\nIn this paper, we study the regularity\nof powers of edge ideals associated to finite simple graphs.\n\nLet $G$ be a finite simple graph on the vertex set $\\{x_1, \\ldots,\nx_n\\}$ and $I(G) := (\\{x_ix_j \\mid \\{i,j\\} \\in E(G)\\}) \\subset \\K[x_1,\n\\ldots, x_n]$ be the edge ideal corresponding to the graph $G$.\nIt is known that\n$\\reg(I(G)^q) = 2q + b$ for some $b$ and $q \\geq q_0$. \nThere are very few classes of graphs for which $b$ and $q_0$ are known. We\nrefer the reader to \\cite{BBH17} and the references cited there for a\nreview of results in the literature in this direction.\nWhile the aim is to obtain the linear polynomial corresponding to\n$\\reg(I(G)^q)$, it seems unlikely that a single combinatorial invariant\nwill represent the constant term for all graphs. This naturally give\nrise to two directions of research. One, to obtain linear polynomials\nfor particular classes of graphs. Two, to obtain upper and\nlower bounds for $\\reg(I(G)^q)$ using combinatorial invariants\nassociated to the graph $G$. It was proved by Beyarslan, H\\`a and Trung that $2q + \\nu(G) -\n1 \\leq \\reg(I(G)^q)$ for all $q \\geq 1$, where $\\nu(G)$ denotes the\ninduced matching number of $G$, \\cite{selvi_ha}. In \\cite{jayanthan},\nthe authors along with Narayanan proved that for a bipartite graph\n$G$, $\\reg(I(G)^q) \\leq 2q + \\cochord(G) - 1$ for all $q \\geq 1$,\nwhere $\\cochord(G)$ denote the co-chordal cover number of $G$. \nThere is no general upper bound known for powers of edge ideals of\narbitrary graphs. Therefore, one may ask:\n\\begin{enumerate}\n \\item[Q1.] Does there exist a graph invariant of $G$, say $\\rho(G)$, such\n   that $\\reg(I(G)^q) \\leq 2q + \\rho(G)$ for all $q \\geq 1$?\n \\item[Q2.] Can one obtain the linear polynomial corresponding to\n   $\\reg(I(G)^q)$ for various classes of graphs?\n\\end{enumerate}\nThis paper evolves around these two questions.\n\nThe first main result of the paper answers Question Q1. \nH\\`a and Woodroofe\n\\cite{HaWood} defined an invariant in terms of star\npacking, denoted by $\\zeta(G)$ (see Section \\ref{reg-pow} for\ndefinition), and proved that $\\reg(I(G)) \\leq \\zeta(G) + 1$.  In this\npaper, we extend H\\'a and Woodroofe's bound to include all powers of\n$I(G)$. We prove:\n\\vskip 2mm \\noindent\n\\textbf{Theorem \\ref{main-first}.}\n{\\em If $G$ is a graph, then for all $q \\geq 1$,}\n  $$\\reg(I(G)^q) \\leq 2q+\\zeta(G)-1.$$\n  \n  \nSo far, in the literature, for the\nclasses of graphs for which the regularity of powers of edge ideals\nhave been computed, they satisfy either $\\reg(I(G)^q) = 2q + \\nu(G) -\n1$ or $\\reg(I(G)^q) = 2q + \\cochord(G) - 1$, for all $q \\geq 2$. In \\cite{jayanthan}, the\nauthors raised the question whether there exists a graph $G$ with\n\\[2q + \\nu(G) - 1 < \\reg(I(G)^q) < 2q + \\cochord(G) - 1, \\text{ for $q \\gg 0$.}\\] As a\nconsequence of our investigation, we obtain a class of graphs which\nattain the upper bound in Theorem \\ref{main-first} and the above\nstrict inequalities are satisfied.\n\nAnother way of bounding the function $\\reg(I(G)^q)$, than using\ncombinatorial invariants, is to relate it to the regularity of $G$\nitself. It was conjectured by Alilooee, Banerjee, Beyarslan and H\\`a,\n\\cite[Conjecture 7.11(2)]{BBH17}: \n\n\\begin{conj}\\label{ABBH-conj}\nIf $G$ is a graph, then for all $q \\geq 1$,\n$\\reg(I(G)^q) \\leq 2q + \\reg(I(G)) - 2.$\n\\end{conj}\nThere are some classes of graphs for which this conjecture is known to\nbe true, see \\cite{BBH17}. As a consequence of the techniques that we have\ndeveloped, we prove the conjecture with an additional hypothesis:\n\\vskip 2mm \\noindent\n\\textbf{Theorem \\ref{weak-conj1}.} {\\em\nLet $G$ be a graph. If every induced subgraph $H$ of \n$G$ has a vertex $x$ with $\\reg(I(H \\setminus N_H[x]))+1 \\leq\n\\reg(I(H))$, then for all $q \\geq 1$,}\n\\[\n \\reg(I(G)^q) \\leq 2q+\\reg(I(G))-2.\n\\]\n\n\nWe then move on to study regularity of powers of vertex decomposable graphs.\nA graph $G$ is said to be vertex decomposable if $\\Delta(G)$ is a vertex decomposable, \nwhere $\\Delta(G)$ denotes the independence complex of $G$ \n(see Section \\ref{reg-vertex} for definition).\nVertex decomposability was first introduced by Provan and Billera \\cite{ProvLouis},\nin the case when all the maximal faces are of equal cardinality, and\nextended to the arbitrary case by Bj\\\"orner and Wachs \\cite{BjWachs}.\nWe have the chain of implications:\n\\begin{equation*}\n \\text{vertex decomposable}\\Longrightarrow \\text{shellable} \\Longrightarrow \\text{sequentially\n Cohen-Macaulay,}\n\\end{equation*} \nwhere a graph $G$ is shellable if $\\Delta(G)$ is a shellable\nsimplicial complex and $G$ is sequentially Cohen-Macaulay if $R/I(G)$\nis sequentially Cohen-Macaulay.\nBoth the above implications are known to be strict.\nRecently, a number of authors have\nbeen interested in classifying or identifying vertex decomposable graphs $G$ in\nterms of the combinatorial properties of $G$, \\cite{BFH15, DochEng09, Wood2}. \nWe prove the Conjecture \\ref{ABBH-conj}, for vertex decomposable graphs.\n\\vskip 2mm \\noindent\n\\textbf{Theorem \\ref{upper-ver}.}\n{\\em Let $G$ be a vertex decomposable graph. Then for all\n$q \\geq 1$,} \\[\n             \\reg(I(G)^q) \\leq 2q+\\reg(I(G))-2.\n            \\] \nAs a consequence, we obtain the linear polynomial corresponding to\n$\\reg(I(G)^s)$ for several classes of graphs such as \n$C_5$-free vertex decomposable, chordal, sequentially Cohen-Macaulay bipartite\ngraphs and certain whiskered graphs.\n\n\nOur paper is organized as follows. In the Section \\ref{pre}, we collect the terminology\nand preliminary results that are essential for the rest of the paper. \nWe prove, in Section \\ref{tech}, several technical lemmas which are\nneeded for the proof of our main results which appear in Sections \\ref{reg-pow} and\n\\ref{reg-vertex}.\n\n\\section{Preliminaries}\\label{pre}\n\nThroughout this article, $G$ denotes a finite simple graph without\nisolated vertices.\nFor a graph $G$, $V(G)$ and $E(G)$ denote the set of all\nvertices and the set of all edges of $G$ respectively.\nThe \\textit{degree} of a vertex $x \\in V(G),$ denoted by \n$\\deg_G(x),$ is the number of edges incident to $x.$\nA subgraph $H \\subseteq G$  is called \\textit{induced} if for $u, v\n\\in V(H)$, $\\{u,v\\} \\in\nE(H)$ if and only if $\\{u,v\\} \\in E(G)$.\nFor $\\{u_1,\\ldots,u_r\\}  \\subseteq V(G)$, let $N_G(u_1,\\ldots,u_r) = \\{v \\in V (G)\\mid \\{u_i, v\\} \\in E(G)~ \n\\text{for some $1 \\leq i \\leq r$}\\}$\nand $N_G[u_1,\\ldots,u_r]= N_G(u_1,\\ldots,u_r) \\cup \\{u_1,\\ldots,u_r\\}$. \nFor $U \\subseteq V(G)$, denote by $G \\setminus U$ \nthe induced subgraph of $G$ on the vertex set $V(G) \\setminus U$.\n\n\nA subset $X$ of $V(G)$ is called \\textit{independent} if there is no edge $\\{x,y\\} \\in E(G)$ \nfor $x, y \\in X$. \nA \\textit{matching} in a graph $G$ is a subgraph consisting of pairwise disjoint edges. \nThe largest size of a   matching in $G$ is called its matching number.\nIf the subgraph is an induced subgraph, the matching is an \\textit{induced matching}. The largest size of an induced matching in $G$ is called its induced \nmatching number and denoted by $\\nu(G)$. \n\nThe \\textit{complement} of a graph $G$, denoted by $G^c$, is the graph on the same\nvertex set in which $\\{u,v\\}$ is an edge of $G^c$ if and only if it is not an edge of $G$.\nA graph $G$ is \\textit{chordal} if every induced\ncycle in $G$ has length $3$, and is co-chordal if the complement graph $G^c$ is chordal.\nThe co-chordal cover number, denoted $\\cochord(G)$, is the\nminimum number $n$ such that there exist co-chordal subgraphs\n$H_1,\\ldots, H_n$ of $G$ with $E(G) = \\bigcup_{i=1}^n E(H_i)$.\n\nOne important tool in the study of regularity of powers of edge ideals\nis even-connections.\nWe recall the concept of \\textit{even-connectedness} from \\cite{banerjee}. \n\n\\begin{definition}\\label{even_connected} Let $G$ be a graph. Two vertices $u$ and $v$ ($u$ may be same as $v$) are said to be even-connected with respect to an $s$-fold \nproducts $e_1\\cdots e_s$, where $e_i$'s are edges of $G$, not necessarily distinct, if there is a path $p_0p_1\\cdots p_{2k+1}$, $k\\geq 1$ in $G$ such that:\n\\begin{enumerate}\n \\item $p_0=u,p_{2k+1}=v.$\n \\item For all $0 \\leq l \\leq k-1,$ $p_{2l+1}p_{2l+2}=e_i$ for some $i$.\n \\item For all $i$, $ \\mid\\{l \\geq 0 \\mid p_{2l+1}p_{2l+2}=e_i \\}\\mid\n   ~ \\leq  ~ \\mid \\{j \\mid e_j=e_i\\} \\mid$.\n \\item For all $0 \\leq r \\leq 2k$, $p_rp_{r+1}$ is an edge in $G$.\n\\end{enumerate}\n\\end{definition}\n\n\\begin{remark} For convenience, we set an edge to be trivially\n  even-connected, i.e., we get the even-connection by setting $k = 0$\n  in the above definition.\n\\end{remark}\nThe following theorem due to Banerjee is used repeatedly throughout this paper:\n\\begin{theorem}\\label{even_connec_equivalent}\\cite[Theorem 6.1 and Theorem 6.7]{banerjee} Let $G$ be a graph with edge ideal\n$I = I(G)$, and let $s \\geq 1$ be an integer. Let $M$ be a minimal generator of $I^s$.\nThen $(I^{s+1} : M)$ is minimally generated by monomials of degree 2, and $uv$ ($u$ and $v$ may\nbe the same) is a minimal generator of $(I^{s+1} : M )$ if and only if either $\\{u, v\\} \\in E(G) $ or $u$ and $v$ are even-connected with respect to $M$.\n \\end{theorem}\n\nPolarization is a process that creates a squarefree monomial ideal (in\na possibly different polynomial ring) from a given monomial ideal,\n\\cite[Section 1.6]{Herzog'sBook}.\nIn this paper, we repeatedly use one of the important properties of the\npolarization, namely:\n\\begin{corollary} \\cite[Corollary 1.6.3(a)]{Herzog'sBook}\\label{pol_reg} Let $I$ be a monomial \nideal in $\\K[x_1, \\ldots, x_n].$ Then\n$$\\reg(I)=\\reg(\\widetilde{I}).$$\n\\end{corollary}\n\n\\section{Technical Lemmas}\\label{tech}\n\nIn this section, we prove several technical results concerning the\ngraph associated with $(\\widetilde{I(G)^{s+1} : e_1 \\cdots e_s})$ and\nsome of its induced subgraphs. We begin by fixing the notation for the\nmost of our results. \n\n\\begin{notation}\\label{esetup}\nLet $G$ be a graph with $V(G)=\\{x_1,\\ldots,x_n\\}$ and $e_1,\\ldots,e_s$, $s \\geq 1,$ be some edges of $G$ which are not\nnecessarily distinct. \nBy Theorem \\ref{even_connec_equivalent} and Corollary \\ref{pol_reg}, $\\widetilde{(I(G)^{s+1}:e_1 \\cdots e_s)}$ is a\nquadratic squarefree monomial ideal in an appropriate polynomial ring. Let\n$G'$ be the graph associated to $\\widetilde{(I(G)^{s+1}:e_1 \\cdots e_s)}$.\n\\end{notation}\n\nOne of the key ingredients in the proof of the main results is a new\ngraph, $G'$, obtained from a given graph $G$ by joining even-connected\nvertices by an edge. Our main aim in this section is to get an upper\nbound for regularity of certain induced subgraphs of $G'$ which in\nturn will help us in bounding $\\reg(I(G'))$. For this purpose, we need\nto understand the structure of the graph $G'$ in more detail. First we\nshow that whiskers can be ignored while taking even-connections.\n\\begin{lemma}\\label{even_obs2}\n Let $G$ be a graph with $e_1,\\ldots,e_s \\in E(G)$, $s \\geq 1$ and $N_G(x)=\\{y\\}$. If $e_i=\\{x,y\\}$, for some $1 \\leq i \\leq s$,\n then $$(I(G)^{s+1}:e_1 \\cdots e_s)=(I(G)^s:\\prod_{j\\neq i}e_j)$$\n\\end{lemma}\n\\begin{proof}\nClearly $(I(G)^s:\\prod_{j\\neq i}e_j) \\subseteq (I(G)^{s+1}:e_1 \\cdots e_s)$.\nLet $uv \\in (I(G)^{s+1}:e_1 \\cdots e_s)$. For $k \\geq 0$, let $(u=p_0) \\cdots (p_{2k+1}=v)$\nbe an even-connection in $G$ with respect to  $e_1 \\cdots e_s$. \nFor $0 \\leq r \\leq k-1$, set $e_{i_r}=\\{p_{2r+1},p_{2r+2}\\}$. If $k=0$, then \n$uv \\in (I(G)^s:\\prod_{j\\neq i}e_j)$. Assume $k\\geq 1$.\nIf $e_i \\neq e_{i_r}$, for all $0 \\leq r \\leq k-1$, then $u$ is an even-connected to $v$ in $G$\nwith respect to $e_1 \\cdots e_{i-1}e_{i+1} \\cdots e_s$. Suppose $e_i=e_{i_r}$, for some\n$0 \\leq r \\leq k-1$. Since $N_G(x)=\\{y\\}$, $p_{2r+1}=p_{2r+3}$. Then\n$(u=p_0)p_1 \\cdots (p_{2r+1}=p_{2r+3})p_{2r+4} \\cdots (p_{2k+1}=v)$ gives an even-connection\nin $G$ with respect to $e_1 \\cdots e_{i-1}e_{i+1} \\cdots e_s$.\nHence $uv \\in (I(G)^s:\\prod_{j\\neq i}e_j)$.\n\\end{proof}\n\nThe following result shows that if a vertex has no intersection with a\nset of edges, then removing such a vertex and taking even-connection\nwith respect to the set of those edges commute with each other.\n\n\n\\begin{lemma}\\label{single_vertex} Let the notation be as in \\ref{esetup}.\nIf for $x \\in V(G)$, $\\{x\\} \\cap e_i=\\emptyset$, for all $1 \\leq i \\leq s$, then \n$$I(G' \\setminus x)=(\\widetilde{I(G \\setminus x)^{s+1}:e_1 \\cdots e_s}).$$ \n\\end{lemma}\n\\begin{proof}\nClearly $(I(G \\setminus x)^{s+1}:e_1 \\cdots e_s) \\subseteq I(G'\n\\setminus x)$.  Suppose $\\{u,v\\} \\in E(G' \\setminus x)$. \nLet $(u=p_0)p_1 \\cdots p_{2k}(p_{2k+1}=v)$ be an even-connection\nin $G$ with respect to $e_1 \\cdots e_s$.  Since $e_i \\cap\n\\{x\\}=\\emptyset$, for all $1 \\leq i \\leq s$, $u$ is even-connected to\n$v$ in $G \\setminus x$ with respect to $e_1 \\cdots e_s$. \n\\end{proof}\n\n\nThe next result talks about new even-connections made out of a given\neven-connection and the neighbors of some of the vertices in the\neven-connection.\n\\begin{lemma}\\label{even_obs} Let the notation be as in  \\ref{esetup}.  \nSuppose $(u=p_0)p_1 \\cdots p_{2k}(p_{2k+1}=v)$ is an even-connection in $G$ with respect to\n$e_1 \\cdots e_s$, for some $k \\geq 1$. If $\\{w,p_i\\} \\in E(G')$, for some \n$0 \\leq i \\leq 2k+1$, then either $\\{u,w\\} \\in E(G')$ or $\\{v,w\\} \\in E(G')$.\n \\end{lemma}\n\\begin{proof}\nIf $i=0,2k+1$, then we are done. \nAssume that $i=2j+1$, for some $j \\geq 0$.\nLet\n$(w=q_0)q_1 \\cdots (q_{2j+1}=p_i)$ be an even-connection with respect to $e_1 \\cdots e_s$ \nin $G$. If $\\{q_{2\\alpha+1},q_{2\\alpha+2}\\}$ and $\\{p_{2\\beta+1},p_{2 \\beta+2}\\}$ do not have\na common vertex, for \nall $0 \\leq \\alpha \\leq j-1$, $j \\leq \\beta \\leq k-1$, then \n$(w=q_0)q_1 \\cdots (q_{2j+1}=p_i)p_{i+1} \\cdots (p_{2k+1}=v)$\nis an even-connection with respect to $e_1 \\cdots e_s$ in $G$.\nTherefore, $wv \\in I(G')$.\nIf $\\{q_{2\\alpha+1},q_{2\\alpha+2}\\}$ and $\\{p_{2\\beta+1},p_{2 \\beta+2}\\}$ \nhave a common vertex,\nfor some $0 \\leq \\alpha \\leq j-1$, $j \\leq \\beta \\leq k-1$, then\nby \\cite[Lemma 6.13]{banerjee},\n$w$ is  even-connected either to $u$ or to $v$ in $G$. \nTherefore either $wu\\in I(G')$ or $wv \\in I(G')$.\nIf $i = 2j+2$, then proof is similar.\n\\end{proof}\n\n\n\n\nThe following lemma, which throws more light into the structure of\n$G'$, is very useful for the induction process.\n\n\\begin{lemma}\\label{tech_lemma} Let the notation be as in  \\ref{esetup}.\nLet $y \\in V(G)$ and $H=G \\setminus N_G[y]$. \nIf $\\{e_1,\\ldots,e_s\\} \\cap E(H) = \\{e_{i_1},\\ldots,e_{i_t}\\}$ and $H'$ is the \ngraph associated to $\\widetilde{(I(H)^{t+1}:e_{i_1} \\cdots e_{i_t})}$, then \n$G' \\setminus N_{G'}[y]$ is an induced subgraph of $H'$.\nIn particular, $$\\reg(I(G' \\setminus N_{G'}[y])) \\leq \\reg(I(H')).$$\n\\end{lemma}\n\n\\begin{proof} \nLet $\\{u,v\\} \\in E(G' \\setminus N_{G'}[y])$. \nLet $(u=p_0)p_1 \\cdots (p_{2k+1}=v)$\nbe an even-connection in $G$ with respect to $e_1 \\cdots e_s$ for some\n$k \\geq 0$. If $p_i \\in N_{G'}[y]$, for some $0 \\leq i \\leq 2k+1$, then by Lemma \\ref{even_obs}, $y$\nis even-connected either to $u$ or to $v$. This contradicts the\nassumption that $\\{u,v\\} \\in G' \\setminus N_{G'}[y]$. Therefore, for\neach $0 \\leq i \\leq 2k+1$, $p_i \\notin N_{G'}[y]$. Hence $\\{u,v\\} \\in E(H')$, which\nproves that $G' \\setminus N_{G'}[y]$ is a subgraph of $H'$.\nIf $a,b \\in V(G' \\setminus\nN_{G'}[y])$ is such that $\\{a,b\\} \\in E(H)$,\nthen $\\{a,b\\} \\in E(G' \\setminus N_{G'}[y])$. Hence \n$G' \\setminus N_{G'}[y]$ is an induced subgraph of $H$.\nThe assertion on the regularity follows from \\cite[Proposition\n4.1.1]{sean_thesis}.  \n\\end{proof}\n\nIn the following results, we show that the even-connections in a\nparent graph with respect to edges coming from an induced subgraph,\ninduces an even-connection in the induced subgraph.\n\\begin{lemma}\\label{ind-reg}\nLet $G$ be a graph and $H$ be an induced subgraph of $G$. If for any $e_1,\\ldots,e_s \\in E(H)$,\n$s \\geq 1$, then $H'$ is an induced subgraph of $G'$, where\n$H'$ and $G'$ are the graph associated to $\\widetilde{(I(H)^{s+1}:e_1 \\cdots e_s)}$\nand $\\widetilde{(I(G)^{s+1}:e_1 \\cdots e_s)}$ respectively. \nIn particular, $$\\reg(I(H')) \\leq \\reg(I(G')).$$\n\\end{lemma}\n\\begin{proof}\nLet $\\{a,b\\} \\in E(H')$. For some $k \\geq 0$, let $(a=p_0)p_1 \\cdots p_{2k}(p_{2k+1}=b)$ be an\neven-connection in $G$ with respect to $e_1 \\cdots e_s$. Since\n$H$ is an induced subgraph of $G$ and $\\{p_{2r+1},p_{2r+2}\\} \\in E(H)$, for all $0 \\leq r \\leq k-1$,\n$(a=p_0)p_1 \\cdots p_{2k}(p_{2k+1}=b)$ is an even-connection in $G$ with respect to \n$e_1 \\cdots e_s$. Therefore, $\\{a,b\\}\\in E(G')$. \nHence $H'$ is the subgraph of $G'$.\nAs in the previous lemma, it can be seen that the subgraph is an induced subgraph.\n\\end{proof}\n\n\nLet the notation be as in  \\ref{esetup}. For some $1 \\leq i \\leq s$, set $e_i =\n\\{x,y\\}$. We further explore the even-connections between $N_{G'}[y]$ and\n$N_{G'}(x)$. If $(u=p_0)p_1\\cdots p_{2k}(p_{2k+1}=y)$ ($u$ may be equal to $y$) be an even-connection in $G$ with respect \nto  $e_1 \\cdots e_s$, for some $k \\geq 0$, then there are three possibilities:\n\\begin{enumerate}\n  \\item $\\{p_{2\\lambda+1}, p_{2\\lambda+2}\\} \\neq e_i$ for any $0 \\leq\n\t\\lambda \\leq k-1$;\n  \\item There exists $0 \\leq \\lambda \\leq k-1$ with $\\{p_{2\\lambda+1},\n\tp_{2\\lambda+2}\\} = e_i$ and  $p_{2\\lambda+1} = y, p_{2\\lambda+2} =\n\tx$;\n  \\item For some $0 \\leq \\lambda \\leq k-1$, $p_{2\\lambda+1}=x$ and $p_{2\\lambda+2}=y$ whenever\n  $\\{p_{2\\lambda+1},p_{2\\lambda+2}\\}=e_i$;\n\\end{enumerate}\nNote that, fixing an even-connection between $u$ and $y$, $(1), (2)$\nand $(3)$ are mutually exclusive. Let \n\\begin{equation}\\label{nota_nbd}\n\\begin{split}\n  X_1=&~\\Big \\{u \\in N_{G'}[y] \\mid (u=p_0)p_1\\cdots p_{2k}(p_{2k+1}=y)\n  \\text{ satisfies either (1) or (2)}\\Big\\};\\\\\n X_2=&~ \\Big\\{u \\in (N_{G'}[y]) \\setminus X_1 \\mid (u=p_0)p_1\\cdots p_{2k}(p_{2k+1}=y)\n  \\text{ satisfies (3)}\\Big\\}.\n \\end{split}\n\\end{equation}\n\\begin{lemma}\\label{even-lemma} Following the notation set above, let\n  $E(G\\setminus N_G[u,x]) \\cap \\{e_1, \\ldots, e_s\\} = \\{e_{j_1},\n  \\ldots, e_{j_t}\\}$ and $(G \\setminus N_G[u,x])'$ denote the graph\n  associated to $\\widetilde{(I(G \\setminus N_G[u,x])^{t+1}:e_{j_1} \\cdots e_{j_t})}$.\n\\begin{enumerate}\n \\item If $u \\in X_1$, then $G' \\setminus N_{G'}[u] \\text{ is an\n   induced subgraph of } (G \\setminus N_{G}[u,x])'.$  In particular,\n   $$\\reg(I(G' \\setminus N_{G'}[u]))\\leq \\reg(I((G \\setminus N_{G}[u,x])')).$$\n  \\item Set $G_1'=G' \\setminus X_1$. If $u \\in X_2$, then $G_1'\n\t\\setminus N_{G_1'}[u]$ is an induced subgraph of $(G \\setminus\n\tN_{G}[u,x])'$. \n\t$$\\reg(I(G_1'\\setminus N_{G_1'}[u]))\\leq \\reg(I((G \\setminus N_{G}[u,x])')).$$\n\\end{enumerate} \n\\end{lemma}\n\\begin{proof}\n (1) Set $H=G' \\setminus N_{G'}[u]$ and $K=(G \\setminus N_{G}[u,x])'$. Let\n$\\{a,b\\} \\in E(H)$ and $(a=q_0)q_1 \\cdots (q_{2l+1}=b)$ be an\neven-connection in $G$ with respect to $e_1 \\cdots e_s$, for some $l \\geq 0$. We show that\n$\\{a,b\\} \\in E(K)$. Note that by Lemma \\ref{even_obs}, $q_i \\notin N_{G'}[u]$, for all $0 \\leq i \\leq 2l+1$.\nSuppose\n$q_i \\in N_{G}[x]$, for some $0 \\leq i \\leq 2k+1$. Since $u \\in X_1$, then\n$\\{u,q_i\\} \\in E(G')$. By Lemma \\ref{even_obs}, $u$ is an even-connected either to\n$a$ or to $b$ in $G$ with respect to $e_1 \\cdots e_s$. This is a contradiction to\n$\\{a,b\\} \\in E(H)$.\nTherefore, $q_i \\notin N_{G}[x]$, for all $0 \\leq i \\leq 2l+1$. Hence \n$a$ is even-connected to $b$ in $G \\setminus N_G[u,x]$ with respect to\n$e_{j_1} \\cdots e_{j_t}$. Therefore, $\\{a,b\\} \\in E(K)$. \n\\vskip 1mm \\noindent\n$(2)$ Set $H=G_1' \\setminus N_{G_1'}[u]$ and $K=(G \\setminus\nN_{G}[u,x])'$. Let $\\{a,b\\} \\in E(H)$ and $(a=q_0)q_1 \\cdots\n(q_{2l+1}=b)$ be an even-connection in $G$ with respect to $e_1 \\cdots\ne_s$, for some $l \\geq 0$.  If $\\{p_{2\\alpha+1},p_{2\\alpha+2}\\}$ and\n$\\{q_{2\\beta+1},q_{2\\beta+2}\\}$ have a common vertex, for some $0 \\leq\n\\alpha \\leq k-1$, $0 \\leq \\beta \\leq l-1$, then by \\cite[Lemma\n6.13]{banerjee}, $u$ is even-connected either to $a$ or to $b$ in\n$G$ with respect to $e_1 \\cdots e_s$. This is a contradiction to our\nassumption that \n$\\{a,b\\} \\in E(H)$. Therefore, for all $0 \\leq \\alpha \\leq k-1$ and $0\n\\leq \\beta \\leq l-1$, $\\{p_{2\\alpha+1},p_{2\\alpha+2}\\}$ and\n$\\{q_{2\\beta+1},q_{2\\beta+2}\\}$ do not have a common vertex.  Note\nthat by Lemma \\ref{even_obs}, $q_i \\notin N_{G'}[u]$, for all $0 \\leq\ni \\leq 2l+1$.  Suppose $q_i \\in N_{G}[x]$, for some $0 \\leq i \\leq\n2l+1$. Since $u \\in X_2$, there exists $0 \\leq \\lambda \\leq k-1$ with\n$p_{2\\lambda+1}=x$  and $p_{2\\lambda+2}=y$. Therefore,\n$(p_{2k+1}=y)p_{2k} \\cdots p_{2\\lambda+1} p_{2\\lambda+2}q_i$ is an\neven-connection in $G$ with respect to $e_1 \\cdots e_s$.  By Lemma\n\\ref{even_obs}, $y$ is even-connected either to $a$ or to $b$ in $G$\nwith respect to $e_1 \\cdots e_s$. Since\nfor all $0 \\leq \\alpha \\leq k-1$ and $0\n\\leq \\beta \\leq l-1$, $\\{p_{2\\alpha+1},p_{2\\alpha+2}\\}$ and\n$\\{q_{2\\beta+1},q_{2\\beta+2}\\}$ do not have a common vertex,\neither $a \\in X_1$ or $b\n\\in X_1$.  This is a contradiction to our assumption that $\\{a,b\\} \\in E(H)$.  Hence $q_i\n\\notin N_G[x]$, for all $0 \\leq i \\leq 2l+1$ so that $\\{a,b\\}$ is an\neven-connection in $G \\setminus N_{G}[u,x]$ with respect to\n$e_{j_1}\\cdots e_{j_t}$. Therefore $\\{a,b\\} \\in E(K)$.\n\nAs in the previous lemma, it can be seen that the subgraphs\nconsidered in (1) and (2) are induced subgraphs.\nThe assertion on the regularity in (1) and (2) follows from \\cite[Proposition\n4.1.1]{sean_thesis}.  \n\\end{proof}\n\n\\section{Regularity powers of graphs}\\label{reg-pow}\n\nIn this section, we obtain a general upper bound for the regularity\npowers of edge ideals of graphs. We first recall the\ndefinition of the invariant $\\zeta(G)$, introduced in \\cite{HaWood}.\n\nLet $v_1 \\in V(G)$ be a vertex of degree at least $2$. If $G \\setminus\nN_G[v_1]$ does not have any vertex of degree at least $2$, then set\n$\\sigma = \\{v_1\\}$ and $H_1$ be the graph obtained by removing all the\nisolated vertices of $G \\setminus N_G[v_1]$. If there exists $v_2 \\in\nV(H_1)$ such that $\\deg_{H_1}(v_2) \\geq 2$, then set $\\sigma = \\{v_1,\nv_2\\}$ and $H_2$ be the graph obtained by removing the isolated\nvertices of $H_1 \\setminus N_{H_1}[v_2]$. Continuing like this, we\nobtain $\\sigma = \\{v_1, \\ldots, v_k\\}$ such that $H_{k-1} \\setminus\nN_{H_{k-1}}[v_k]$ is a collection of disconnected edges, say $\\{w_1,\n\\ldots, w_r\\}$ and isolated vertices. Let $\\PP = \\{N_G[v_1],\n\\ldots, N_G[v_k]\\} \\cup \\{w_1, \\ldots, w_r\\}.$ We call this $\\PP$ a\n\\textit{star packing} of $G$ and set $\\zeta_{\\PP}(G) = k+r$. Define $$\\zeta(G)\n:= \\max \\Big\\{\\zeta_{\\PP}(G) \\mid \\PP \\text{ is a star packing of } G \\Big\\}.$$\nFor example, if $G = C_n$, cycle on $n$ vertices, then for any $x \\in\nV(G)$, $N_G[x]$ is a path on $3$ vertices and hence $G \\setminus\nN_G[x]$ is a path on $n-3$ vertices. A maximal star packing can\nbe obtained by successively taking out $N_G[x]$, where $x$ is the \nneighbour of a degree $1$ vertex. Therefore, if $n \\equiv \\{0,1\\}(\\text{mod }3)$,\nthen $\\zeta(G) = \\lfloor \\frac{n}{3} \\rfloor = \\nu(G)$ and if $n\n\\equiv 2(\\text{mod }3)$, then $\\zeta(G) = \\lfloor \\frac{n}{3} \\rfloor + 1 =\n\\nu(G) + 1$.\n\n\nIt may be noted that for a graph $G$, $\\nu(G) \\leq \\zeta(G)$,\n\\cite{HaWood}. H\\`a and Woodroofe proved:\n\\begin{theorem}\\cite[Theorem 1.6]{HaWood} \\label{reg-upper} Let $G$ be\n a graph. Then $$\\reg(I(G)) \\leq \\zeta(G)+1.$$ \n\\end{theorem}\n\nIt is easy to see that $\\zeta(G)$ is at most the matching number of\n$G$.  There are two general upper bounds known for the class of edge\nideals, namely, $\\reg(I(G)) \\leq \\cochord(G) + 1$, \\cite[Theorem\n1]{russ} and $\\reg(I(G)) \\leq \\text{min-max}(G) + 1$, \\cite{ha_adam,\nruss}, where min-max$(G)$ denotes the minimum number of a maximal\nmatching in $G$.  We would like to note here that the invariants\n$\\zeta(G)$, $\\cochord(G)$ and min-max$(G)$ are not comparable in\ngeneral, as can be seen from the following examples.\n\\begin{example}\\label{ex:incomparable}\nLet $G=C_7$. Then one can see that the co-chordal subgraphs of\n$G$ are paths with at most $3$ edges so that $\\cochord(G)=3$. Also, one seen that \n$\\zeta(G)=2$. Let $H$ be obtained from $C_4$ by attaching a pendant to\nany of the vertices. \nIt is easy to see that $\\cochord(H)=1$ and $\\zeta(H)=2$.\n\nLet $G$ be the graph obtained by adding a pendant vertex each to two\nvertices having a common neighbor in $C_6$. Then min-max$(G) = 2$ and\n$\\zeta(G) = 3$. If $H = C_4$, then it can be seen that min-max$(G) =\n2$ while $\\zeta(G) = 1$.\n\\end{example}\n\nWe now make an observation about the behaviour of the invariant\nwhich is crucial in our inductive arguments.\n\n\\begin{obs}\\label{inv-obs}\nLet $x$ be a vertex of $G$ of degree at least $2$, then $\\zeta(G \\setminus\nN_G[x]) + 1 \\leq \\zeta(G)$.\n\\end{obs}\n\\begin{proof}\nLet $H = G \\setminus N_G[x]$. Let $\\PP'$ be a star packing of $H$ such\nthat $\\zeta_{\\PP'}(H) = \\zeta(H)$. Then $\\PP = \\PP' \\cup \\{N_G[x]\\}$\nis a star packing of $G$. Thus, \n$\\zeta(H) + 1 = \\zeta_{\\PP'}(H) + 1 = \\zeta_{\\PP}(G) \\leq \\zeta(G).$\n\\end{proof}\n\nWe first prove that for a given graph $G$ and edges $e_1, \\ldots,\ne_s$, the regularity of $G'$\nis bounded above by one more than $\\zeta(G)$.\n\n\n\\begin{theorem}\\label{base-reg}\n If $G$ is a  graph, then for any \n $e_1,\\ldots,e_s \\in E(G)$, $s \\geq 1$\n$$\\reg(I(G)^{s+1}:e_1 \\cdots e_s) \\leq  \\zeta(G)+1.$$\n\\end{theorem}\n\\begin{proof}\nLet $G'$ be the graph associated to \n$\\widetilde{(I(G)^{s+1}:e_1 \\cdots e_s)}$ contained in an appropriate polynomial ring \n$R_1$.\n  We prove the assertion by induction\non $s$. Suppose $s=1$.  \nSet $e_1=\\{x,y\\}$. If either $\\deg_G(x)=1$ or $\\deg_G(y)=1$, then \n  by Lemma \\ref{even_obs2}, $I(G')=I(G)$. Therefore by \n  Theorem \\ref{reg-upper}, $\\reg(I(G'))=\\reg(I(G))\\leq \\zeta(G)+1$.\n\nSuppose $\\deg_G(x)>1$ and $\\deg_G(y)>1$. Setting\n$U=N_{G'}(y)=\\{y_1,\\ldots,y_r\\}$ and\n $J=I(G')$, consider the exact sequences:\n\\begin{eqnarray}\\label{main-exact-seq}\n  0 & \\longrightarrow & \\frac{R_1}{(J : y_1)}(-1)\n  \\overset{\\cdot y_1}{\\longrightarrow} \\frac{R_1}{J} \\longrightarrow\n  \\frac{R_1}{(J , y_1)} \\longrightarrow 0;\\nonumber \\\\\n \n \n \n& & \\hspace*{1cm} \\vdots\\hspace*{5cm}\\vdots \\hspace*{3cm}\\vdots \\\\ \n0 & \\longrightarrow & \\frac{R_1}{((J,\n  y_1,\\ldots,y_{r-1}):y_r)}(-1) \n\\overset{\\cdot y_r}{\\longrightarrow}\n\\frac{R_1}{(J, y_1,\\ldots,y_{r-1})} \\longrightarrow\n\\frac{R_1}{(J, U)} \\longrightarrow 0.\\nonumber\n\\end{eqnarray}\n It follows from these exact sequences that\n\\[\n  \\reg(R_1/J)  \\leq  \\max \\left\\{\n\t\\begin{array}{l}\n\t  \\reg\\left(\\frac{R_1}{(J :y_1)}\\right)+1,\\ldots, \n\t\t\\reg\\left(\\frac{R_1}{(J,y_1,\\ldots,y_{r-1}):y_r)}\\right) + 1,~ \n\t\t\\reg\\left(\\frac{R_1}{(J,~U)}\\right)\n\t  \\end{array}\\right\\}.\n\\]\nWe now prove that each of the regularities appearing on the right hand\nside of the above inequality is bounded above by $\\zeta(G)$. \n \nSince $(J,U)$ corresponds to an induced subgraph of $G$, by \\cite[Proposition\n4.1.1]{sean_thesis} and Theorem \\ref{reg-upper},\n\\begin{equation*}\\label{case2:eq3}\n\\reg \\left(\\frac{R_1}{(J,U)}\\right) \\leq \\reg \\left(\\frac{R}{I(G)}\\right)\\leq \\zeta(G).\n \\end{equation*}\nLet the notation be as in Equation (\\ref{nota_nbd}).\nSince $N_{G'}(y)=N_G(y)$, $U \\subseteq X_1$. \nNote that $x \\in N_G(y)$ and $e_1 \\cap E(G \\setminus N_{G}[y_j,x])=\\emptyset$, for any $y_j \\in U$. Therefore,\nwe have\n$$\n\\begin{array}{lcll}\n   \\reg((J:y_j)) & = &\\reg(I(G' \\setminus N_{G'}[y_j])) \\leq  \n   \\reg(I(G \\setminus N_{G}[y_j,x])) & \\mbox{ (by Lemma \\ref{even-lemma})}\\\\ \n  \n   & \\leq & \\reg(I(G \\setminus N_{G}[x]))\\leq\\zeta(G \\setminus N_{G}[x])+1 \\leq \\zeta(G). & \\mbox{ (by Observation \n   \\ref{inv-obs}) }\\\\\n   \\end{array}\n$$\nSince $((J,y_1,\\ldots,y_{j-1}):y_j)$\ncorresponds to an induced subgraph of $(J:y_j)$, it follows that\n$$\\reg \\left( \\frac{R_1}{((J,y_1,\\ldots,y_{j-1}):y_j)}\\right)+1 \\leq \n\\reg \\left( \\frac{R_1}{(J:y_j)}\\right) +1\\leq \\zeta(G).$$\n\nTherefore, \n$\\reg\\left( \\frac{R_1}{J}\\right) \\leq \\zeta(G).$\n\nSuppose $s > 1$. Assume by induction that for any graph $G$ and for any $e_1,\n\\ldots, e_{s-1}\\in E(G)$, $\\reg(I(G)^s : e_1 \\cdots e_{s-1}) \\leq\n\\zeta(G) + 1$.\n\nLet $G$ be a graph, $e_1, \\ldots, e_s \\in E(G), ~s\\geq 2$.\nFor $1 \\leq i \\leq s$, let $e_i=\\{a_i,b_i\\}$. If for some $1 \\leq j \\leq s$, either\n$\\deg_G(a_j)=1$ or $\\deg_G(b_j)=1$, then\nby Lemma \\ref{even_obs2},\n$$\\reg((I(G)^{s+1}:e_1 \\cdots e_s))=\\reg((I(G)^{s}:e_1 \\cdots e_{j-1}e_{j+1} \\cdots e_{s})) \n\\leq \\zeta(G)+1,$$\nwhere the last inequality follows by induction hypothesis on $s$. \n\nSuppose\n$\\deg_G(a_i)>1$ and $\\deg_G(b_i)>1$, for all $1 \\leq i \\leq s$.\nLet $N_{G'}(b_s)=\\{y_1,\\ldots,y_p,z_1,\\ldots,z_q\\}$.  \nFollowing the notation in Equation (\\ref{nota_nbd}),\nset $X_1=\\{y_1,\\ldots,y_p\\}$,\n$X_2=\\{z_1,\\ldots,z_q\\}$ and $J=I(G')$. \nIt follows from set of short exact sequences, similar to Equation\n(\\ref{main-exact-seq}), that\n\\[\n  \\reg(R_1/J)  \\leq  \\max \\left\\{\n\t\\begin{array}{l}\n\t  \\reg\\left(\\frac{R_1}{(J :y_1)}\\right)+1, \\ldots, \\reg\\left(\\frac{R_1}{(J,X_1) : z_1}\\right)+1,  \n\t\t\\ldots, \\\\\t\t\n\t\t\\;\\; \\\\\n\t\t\\reg\\left(\\frac{R_1}{((J,X_1,z_1,\\ldots,z_{q-1}):z_q)}\\right) + 1,~ \n\t\t\\reg\\left(\\frac{R_1}{(J,X_1,X_2)}\\right).\n\\end{array}\\right.\n\\]\nNow,\n\\[\n  \\begin{array}{llll}\n\\reg((J,X_1,X_2))& = & \\reg(I(G' \\setminus N_{G'}[b_s])) &\n(\\text{by \\cite[Remark 2.5]{selvi_ha}}) \\\\\n & \\leq&  \\reg(I((G \\setminus\n N_{G}[b_s])')) & (\\text{by Lemma \\ref{tech_lemma}}),\n \\end{array}\\]\n where  \n $E(G \\setminus N_G[b_s]) \\cap\n \\{e_1,\\ldots,e_s\\}=\\{e_{j_1},\\ldots,e_{j_t}\\}$ and $(G \\setminus N_{G}[b_s])'$\n is the graph associated to $\\widetilde{(I(G \\setminus\n   N_G[b_s])^{t+1}:e_{j_1} \\cdots e_{j_t})}$.   \n   Since $a_s \\in N_G(b_s), e_s \\notin E(G\\setminus N_G[b_s])$ so that\n   $t<s$. Hence by\n induction hypothesis on $s$ and Observation \\ref{inv-obs}, we get\n  $$\\reg(I(G' \\setminus N_{G'}[b_s])) \\leq \\reg(I((G \\setminus N_{G}[b_s])')) \\leq \n  \\zeta(G \\setminus N_{G}[b_s])+1 \\leq \\zeta(G).$$\nLet $E(G \\setminus N_G[y_i,a_s]) \\cap\n\\{e_1,\\ldots,e_s\\}=\\{e_{j_1},\\ldots,e_{j_t}\\}$ and $(G \\setminus\nN_G[y_i,a_s])'$ be the graph associated to \n  $\\widetilde{(I(G \\setminus N_G[y_i,a_s])^{t+1}:e_{j_1} \\cdots\n  e_{j_t})}$.\n  For any $y_i \\in X_1$, we have\n  \\[\n\t\\begin{array}{llll}\n\t  \\reg((J:y_i))=\\reg(I(G' \\setminus N_{G'}[y_i])) & \\leq &\n\t  \\reg(I((G \\setminus N_G[y_i,a_s])')) & (\\text{by Lemma\n\t\t\\ref{even-lemma}}) \\\\\t\t\n\t  & \\leq & \\zeta(G \\setminus N_{G}[y_i,a_s])+1 & (\\text{by\n\t  induction on } s) \\\\\n\t  & \\leq & \\zeta(G \\setminus N_{G}[a_s])+1 \n\t   \\leq  \\zeta(G). & (\\text{by Observation \\ref{inv-obs}})\n\t\\end{array}\n  \\]\nSince $((J,y_1,\\ldots,y_{i-1}):y_i)$\ncorresponds to an induced subgraph of $(J:y_i)$, it follows that\n$$\\reg(((J,y_1,\\ldots,y_{i-1}):y_i)) \\leq \\reg(J:y_i)\\leq \\zeta(G).$$\n \nLet $G_1'=G' \\setminus X_1$. For any $z_i \\in X_2$, we may conclude,\nas done earlier, that \n\\[\\reg((J,X_1:z_i)) \\leq \\reg(I((G \\setminus N_G[z_i,a_s])'))\n\\leq \\zeta(G \\setminus N_{G}[a_s])+1\\leq \\zeta(G).\\]\n \nSince $((J,X_1,z_1,\\ldots,z_{i-1}):z_i)$\ncorresponds to an induced subgraph of $(J,X_1:z_i)$, it follows that\n$$\\reg(((J,X_1,z_1,\\ldots,z_{i-1}):z_i)) \\leq \\reg(J,X_1:z_i)\\leq \\zeta(G).$$\n\nTherefore\n$\\reg(J) \\leq \\zeta(G)+1.$\nHence, for any $e_1,\\ldots,e_s \\in E(G)$, $s \\geq 1$,\n$$\\reg((I(G)^{s+1}:e_1 \\cdots e_s)) \\leq  \\zeta(G)+1.$$\n \\end{proof}\nNow we prove an upper bound for the regularity of powers of edge\nideals of  graphs. \n \\begin{theorem}\\label{main-first}\n  If $G$ is a graph, then for all $q \\geq 1$,\n  $$\\reg(I(G)^q) \\leq 2q+\\zeta(G)-1$$\n \\end{theorem}\n\\begin{proof}\n We prove by induction on $q$. If $q=1,$ then the\n assertion follows from Theorem \\ref{reg-upper}.\nAssume that $q >1$. By applying \\cite[Theorem 5.2]{banerjee} and using\ninduction, it is enough to prove that for edges $e_1, \\ldots, e_q$ of\n$G$, \n$\\reg(I(G)^{q+1}:e_1 \\cdots e_q) \\leq \\zeta(G))+1$ for all $q > 1$. \nThis follows from  Theorem \\ref{base-reg}.\n\\end{proof}\n\n\n\nIf $G_1$ and $G_2$ are graphs for which the linearity of $\\reg(I(G_1)^s)$ and\n$\\reg(I(G_2)^s)$ are known for $s \\geq s_1$ and $s \\geq s_2$\nrespectively, then by \\cite[Theorem 5.7]{nguyen_vu}, it is known that\n$\\reg(I(G_1 \\coprod G_2)^s)$ is linear for $s \\geq s_1+s_2$. Using\nthis result, we obtain a class of graphs for which the upper bound in\nTheorem \\ref{main-first} is attained.\n\n\\begin{proposition}\\label{disjoint}\n For $p \\geq 0,r > p$, let\n$\n  H=\\left(\\coprod_{i=1}^p C_{n_i}\\right) \\coprod \\left(\\coprod_{j=p+1}^rC_{n_j}\\right), \n $ where \n$n_1,\\ldots,n_p \\equiv 2(mod~3)$, $n_{p+1},\\ldots,n_r \\equiv\n\\{0,1\\}(mod~3)$.\nThen for all $q \\geq 1$,\n$$\\reg(I(H)^q)=2q+\\zeta(H)-1.$$ \n\\end{proposition}\n\\begin{proof}\n It follows from \n\\cite[Theorem 7.6.28]{sean_thesis} and \\cite[Theorem 5.2]{selvi_ha}\nthat\n\\[\n  \\reg(I(C_n)) = \\left\\{\\def\\arraystretch{1.2}%\n  \\begin{array}{@{}c@{\\quad}l@{}}\n    \\nu(C_n)+1 & \\text{ if $n \\equiv \\{0,1\\}~(mod~3),$}\\\\\n    \\nu(C_n)+2 & \\text{ if $n \\equiv 2~(mod~3),$}\\\\\n  \\end{array}\\right.\n\\]\nand that for all $q \\geq 2$, $\\reg(I(C_n)^q)=2q+\\nu(C_n)-1.$\nIf $q=1$, then by \\cite[Lemma 8]{russ}, we get $\\reg(I(H))=\\zeta(H)+1$. If $p = 0$, then\n$H=(\\coprod_{j=p+1}^rC_{n_j})$.\nBy  \\cite[Theorem 5.7]{nguyen_vu},\n$\\reg(I(H)^q)=2q+\\zeta(H)-1, \\text{for all $q \\geq 1$}.$\n\nSuppose $p>0$. Set $H'=(\\coprod_{j=p+1}^rC_{n_j})$.\nLet $L_1 = H' \\coprod C_{n_1}$, where $n_1 \\equiv 2(\\text{mod }3)$.\nThen  by \\cite[Proposition 2.7]{HTT},\n$\\reg(I(L_1)^2)=\\zeta(L_1)+3.$\nBy \\cite[Theorem 5.7]{nguyen_vu}, for $q \\geq 3$, we have\n$\\reg(I(L_1)^q)=2q+\\zeta(L_1)-1.$\nFor $i \\geq 2$, let $L_i = L_{i-1} \\coprod C_{n_i}$, where $n_i \\equiv\n2(\\text{mod }3)$.\nRecursively applying \\cite[Proposition 2.7]{HTT} and \\cite[Theorem\n5.7]{nguyen_vu}, to the graphs $L_i$, we get for all $q \\geq 2$,\n$\\reg(I(H)^q)=2q+\\zeta(H)-1.$\n\\end{proof}\n\nIn \\cite{jayanthan}, the authors asked if there exists a graph \n$G$ with $2q + \\nu(G) - 1 < \\reg(I(G)^q) < 2q + \\cochord(G) - 1$ for\nall $q \\gg 0$, \\cite[Question 5.8]{jayanthan}. We show that some of\nthe graphs considered in Proposition\n\\ref{disjoint} satisfy this inequality.\nLet $H$ be a graph as in Proposition \\ref{disjoint}, with $n_j \\equiv\n1(\\text{mod }3)$ for $j = p+1, \\ldots, q$. Then $\\nu(H) =\n\\sum_{i=1}^q \\lfloor\\frac{n_i}{3} \\rfloor$, $\\zeta(H) =\np+ \\sum_{i=1}^q \\lfloor \\frac{n_i}{3} \\rfloor$ and $\\cochord(G) =\nq+ \\sum_{i=1}^q \\lfloor \\frac{n_i}{3} \\rfloor$. Therefore, we get\n$$2q+\\nu(H)-1<\\reg(I(H)^q) = 2q + \\zeta(H) - 1 <2q+\\cochord(H)-1.$$\n\n\nUsing techniques very similar to the ones used in the proof of Theorem\n\\ref{base-reg}, we prove a weaker version of Conjecture\n\\ref{ABBH-conj}. \n\n\\begin{theorem}\\label{weak-conj1} Let $G$ be a graph. If every induced subgraph $H$ of \n$G$ has a vertex $x$ with $\\reg(I(H \\setminus N_H[x]))+1 \\leq\n\\reg(I(H))$, then for all $q \\geq 1$,\n\\[\n \\reg(I(G)^q) \\leq 2q+\\reg(I(G))-2.\n\\]\n\\end{theorem}\n\\begin{proof}\nLet $G$ be a graph satisfying the given hypothesis. We prove the\nassertion by induction on $q$. If $q=1,$ then we are done.\nAssume that $q >1$. \nFor any graph $K$, set \n \\[\n  \\PP(K)=\\{x \\in V(K) \\mid \\reg(I(K)) \\geq \\reg(I(K \\setminus N_K[x]))+1\\}.\n   \\]\nBy hypothesis, $\\PP(G) \\neq \\emptyset$.\nBy applying \\cite[Theorem 5.2]{banerjee} and using\ninduction, it is enough to prove that for edges $e_1, \\ldots, e_s$ of\n$G$,\n$\\reg((I(G)^{s+1}:e_1 \\cdots e_s)) \\leq \\reg(I(G))$ for all $s \\geq 1$. \nWe prove this by induction\non $s$.\n\n\nLet $G'$ be the graph associated to the ideal\n  $\\widetilde{(I(G)^{s+1}:e_1\\cdots e_s)}$ which is contained in an appropriate polynomial ring \n  $R_1$.\nSuppose $s=1$. \n \\vskip 1mm \\noindent\n\\textsc{Case 1:} Suppose $e_1\\cap \\PP(G) \\neq \\emptyset$.\n\\vskip 1mm \\noindent \nLet $e_1=\\{x,y\\}$ with $x \\in \\PP(G)$.\nWe proceed as in the proof of Theorem \\ref{base-reg}. \n If either $\\deg_G(x)=1$ or $\\deg_G(y)=1$, \nthen by Lemma \\ref{even_obs2}, $I(G')=I(G)$. Therefore,\n$\\reg(I(G'))=\\reg(I(G))$.\n\\vskip 1mm \\noindent\nSuppose  $\\deg_G(x)>1$ and $\\deg_G(y)>1$. Setting \n   $U=N_G(y)=\\{y_1,\\ldots,y_r\\}$ and \n $J=I(G')$ and considering the exact sequences as in Equation\n (\\ref{main-exact-seq}), we get\n \\[\n  \\reg(R_1/J)  \\leq  \\max \\left\\{\n\t\\begin{array}{l}\n\t  \\reg\\left(\\frac{R_1}{(J :y_1)}\\right)+1, \n\t\t\\ldots, \n\t\t\\reg\\left(\\frac{R_1}{(J,y_1,y_2,\\ldots,y_{r-1}):y_r)}\\right) + 1,~ \n\t\t\\reg\\left(\\frac{R_1}{(J,~U)}\\right).\n\\end{array}\\right.\n\\]\nProceeding as in that proof, we conclude that $\\reg(R_1/(J,U)) \\leq\n\\reg(R/I(G))$ and $\\reg(J: y_j) \\leq \\reg(I(G \\setminus N_G[x])) <\n\\reg(I(G))$, where the last inequality follows from the hypothesis. \nSince $((J,y_1,\\ldots, y_{j-1}): y_j)$ corresponds to\nan induced subgraph of $(J : y_j)$, we get\n$$\\reg \\left( \\frac{R_1}{((J,y_1,\\ldots,y_{j-1}):y_j)}\\right)+1 \\leq \n\\reg \\left( \\frac{R_1}{(J:y_j)}\\right) +1\\leq \\reg\\left(\\frac{R}{I(G)}\\right).$$\nTherefore, \n$\\reg\\left( \\frac{R_1}{J}\\right) \\leq \\reg(\\frac{R}{I(G)}).$\n\\vskip 1mm \\noindent\n\\textsc{Case 2:} Suppose $e_1 \\cap \\PP(G) = \\emptyset$. \n\\vskip 1mm \\noindent  \n Let $|V(G)|=n$. We proceed by induction on $n$. If $n \\leq 4$, then it can be\nseen that $e \\cap \\PP(G)\\neq \\emptyset$, for all $e \\in E(G)$ so that\n\\textsc{Case 2} does not occur.\nTherefore $n \\geq 5$.  \n \\noindent\n \\begin{minipage}{\\linewidth}\n \n\\begin{minipage}{0.2\\linewidth} \n\\begin{figure}[H]\n \\begin{tikzpicture}[scale=.75]\n\\draw (4,3)-- (3,2);\n\\draw (4,3)-- (5,2);\n\\draw (3,2)-- (5,2);\n\\draw (3,2)-- (3.01,0.59);\n\\draw (5,2)-- (4.99,0.61);\n\\draw (3.01,0.59)-- (4.99,0.61);\n\\draw (4.99,0.61)-- (3.01,0.59);\n\\draw (5,2)-- (3,2);\n\\begin{scriptsize}\n\\fill [color=black] (4,3) circle (1.5pt);\n\\draw[color=black] (4.25,3.23) node {$t_1$};\n\\fill [color=black] (3,2) circle (1.5pt);\n\\draw[color=black] (2.88,2.25) node {$t_5$};\n\\fill [color=black] (5,2) circle (1.5pt);\n\\draw[color=black] (5.25,2.23) node {$t_2$};\n\\fill [color=black] (3.01,0.59) circle (1.5pt);\n\\draw[color=black] (2.79,0.7) node {$t_4$};\n\\fill [color=black] (4.99,0.61) circle (1.5pt);\n\\draw[color=black] (5.23,0.85) node {$t_3$};\n\\end{scriptsize}\n\\end{tikzpicture}\n\\end{figure}\n\\end{minipage}\n\\begin{minipage}{0.78\\linewidth}\nSuppose $n=5$.  There are $23$ simple graphs (without isolated\nvertices) on $5$ vertices. Among these graphs, it can be verified,\nmanually or using computational packages such as Macaulay 2 \\cite{M2}\nor SAGE \\cite{sage}, that all, except the graph given on the left,\nsatisfy the property that $G \\setminus \\PP(G)$ is a set of isolated\nvertices. \n\\end{minipage}\n\\end{minipage}\nFor this graph $G$, $\\reg(I(G)) = 2$ and $\\PP(G) = \\{t_2, t_5\\}$. Hence\n$\\{t_3,t_4\\} \\in E(G \\setminus \\PP(G))$ and $(I(G)^2 : t_3t_4) = I(G)$ so\nthat $G' = G$. Therefore, the assertion holds true.\n  \nNow assume that $n>5$. By induction, assume that if $K$ is a\ngraph with $|V(K)| < n$ and for every induced subgraph $K'$ of $K$,\nthere exists $z \\in V(K')$ such that\n$\\reg(I(K' \\setminus N_{K'}[z]))+1 \\leq \\reg(I(K'))$,\nthen $\\reg(I(K)^2 :e) \\leq \\reg(I(K))$ for every\n$e\\in E(K)$ such that $e \\cap \\PP(K)= \\emptyset$.\n\nLet $x \\in \\PP(G)$. Then by \\cite[Theorem 3.4]{Ha2},\n\\[\\reg(I(G')) \\leq \\max\n\\Big\\{\\reg(I(G' \\setminus \\{x\\})), \\reg(I(G' \\setminus\nN_{G'}[x])+1\\Big\\}.\\]\nSince $e_1 \\cap \\{x\\}=\\emptyset$, by Lemma \\ref{single_vertex}, \n$$\\reg(I(G' \\setminus x))=\\reg((I(G \\setminus x)^2:e_1)).$$\nNote that $G \\setminus x$ is a graph with $|V(G \\setminus x)|<n$ and every induced\nsubgraph of $G \\setminus x$ is an induced subgraph of $G$.\nIf $e_1 \\cap \\PP(G \\setminus x) \\neq \\emptyset$, then by Case 1, or if\n$e_1 \\cap \\PP(G \\setminus x) = \\emptyset$, then by induction on the number of vertices, we get\n$$\\reg((I(G \\setminus x)^2:e_1)) \\leq \\reg(I(G \\setminus x)) \\leq \\reg(I(G)).$$\nNow we prove that $\\reg(I(G' \\setminus N_{G'}[x])+1 \\leq \\reg(I(G))$.\nNote that $G \\setminus N_G[x]$ is a graph with $|V(G \\setminus N_G[x])|<n$.\nIf $e_1 \\cap \\PP(G \\setminus N_G[x]) \\neq \\emptyset$, then by Case 1, or if\n$e_1 \\cap \\PP(G \\setminus N_G[x]) = \\emptyset$, then by induction on the number of vertices, we get\n\\[\n\\begin{array}{llll}\n\\reg(I(G' \\setminus N_{G'}[x]))+1 & \\leq & \\reg(I(G \\setminus N_G[x])^2:e_1)+1 & \n(\\text{by Lemma }\\ref{tech_lemma}) \\\\\n\n \n   & \\leq & \\reg(I(G \\setminus N_G[x]))+1 &  \\\\\n   & \\leq & \\reg(I(G)). & \n\\end{array}\n\\]\nHence $\\reg(I(G')) \\leq \\reg(I(G))$. This proves the case $s =1$.\n  \nSuppose $s > 1$. We now show that $\\reg(I(G)^{s+1}: e_1\\cdots e_s)\n\\leq \\reg(I(G))$. Let $e_i = \\{a_i, b_i\\}$ for $1 \\leq i \\leq s$. If\n$\\deg_G(a_i) = 1$ or $\\deg_G(b_i)$ for some $i$, then by Lemma\n\\ref{even_obs2}, it follows that \n\\[\\reg(I(G)^{s+1} : e_1\\cdots e_s) = \\reg(I(G)^s : e_1\\cdots\n  e_{i-1}e_{i+1}\\cdots e_s) \\leq \\reg(I(G)),\\]\nwhere the last inequality follows from induction on $s$.\n\nAssume now that $\\deg_G(a_i) \\geq 2$ and $\\deg_G(b_i) \\geq 2$ for all\n$1 \\leq i \\leq s$.\n  \\vskip 1mm \\noindent\n\\textsc{Case 3:} Suppose $e_i \\cap \\PP(G) \\neq \\emptyset$, for some $1\n\\leq i \\leq s$. \n\\vskip 1mm \\noindent\nWithout loss of generality, assume that $e_s \\cap \\PP(G) \\neq\n\\emptyset$ and $a_s \\in \\PP(G)$.\nProceeding as in the proof Theorem \\ref{base-reg} following the same\nnotation, one gets \n  \\begin{eqnarray*}\n   \\reg(J,X_1,X_2)  &\\leq & \\reg(I(G' \\setminus N_{G'}[b_s]))\\leq \\reg(I(G \\setminus N_G[b_s])) \\leq\n  \\reg(I(G));\\\\\n  \\reg(J : y_i) & \\leq & \\reg(I((G \\setminus N_G[y_i,a_s])')) \\leq \\reg(I((G \\setminus N_G[a_s])')) \\\\\n  & \\leq&  \\reg(I(G \\setminus N_G[a_s])) < \\reg(I(G));\n  \\\\\n  \\reg(J,X_1 : z_i) &\\leq &\\reg(I((G \\setminus N_G[y_i,a_s])')) \\leq \\reg(I((G \\setminus N_G[a_s])'))\\\\\n  & \\leq & \\reg(I(G \\setminus N_G[a_s])) < \\reg(I(G)).\n\\end{eqnarray*}\nFor the above conclusions, we use, in a similar manner as in the proof\nof Theorem \\ref{base-reg}, Lemmas \\ref{tech_lemma}, \\ref{ind-reg},\n\\ref{even-lemma} and induction on $s$. Using these inequalities, we\nconclude that $\\reg(J)\n\\leq \\reg(I(G))$.\n\\vskip 1mm \\noindent\n\\textsc{Case 4:}\nSuppose $e_i \\cap \\PP(G) = \\emptyset$, for all $1 \\leq i \\leq s$.\n\nIf $|V(G)| \\leq 4$, then one can see that the case $e_i \\cap \\PP(G) =\n\\emptyset$ does not occur. If $|V(G)| = 5$, then as remarked in\n\\textsc{Case 2}, one can see that there is only one graph $G$, which\nis given there, such that $G \\setminus \\PP(G)$ has an edge. In this\ncase, the only possibility is $e_i = \\{t_3, t_4\\}$ for all $i = 1,\n\\ldots, s$. Hence $G' = G$ and hence the assertion holds true.\n\nNow assume that $|V(G)| = n > 5$.\nLet $x \\in \\PP(G)$. Then by by \\cite[Theorem 3.4]{Ha2}, \n\\[\\reg(I(G')) \\leq \\max\n\\Big\\{\\reg(I(G' \\setminus \\{x\\})), \\reg(I(G' \\setminus\nN_{G'}[x]))+1\\Big\\}.\\]\n\nBy Lemma \\ref{single_vertex}, $\\reg(I(G' \\setminus x)) =\n\\reg(I(G\\setminus x)^{s+1} : e_1 \\cdots e_s))$ and by Lemma\n\\ref{tech_lemma},  $\\reg(I(G' \\setminus N_{G'}[x])) \\leq \\reg(I(G\n\\setminus N_G[x])^{t+1} : e_{i_1}\\cdots e_{i_t})$, where $E(G\\setminus\nN_G[x]) \\cap \\{e_1, \\ldots, e_s\\} = \\{e_{i_1}, \\ldots, e_{i_t}\\}$.\nProceeding as in \\textsc{Case 2}, using the above inequalities and\ninduction on $|V(G)|$ as well as on $s$, \none can conclude that $\\reg(I(G')) \\leq \\reg(I(G))$. \nTherefore $\\reg(I(G)^{s+1} : e_1\\cdots e_s) \\leq \\reg(I(G))$.\n\\end{proof}\n\nIn view of the above theorem, we would like to ask:\n\n\\begin{question}\nDoes every finite simple graph $G,$ having no isolated vertices, have\na vertex $x$ such that $\\reg(I(G): x) + 1 \\leq \\reg(I(G))$?\n\\end{question}\nWe note here that if the answer to the above question is positive,\nthen it follows from Theorem \\ref{weak-conj1} that the Conjecture\n\\ref{ABBH-conj} is true.\n\n\n\n\\section{Vertex decomposable graphs}\\label{reg-vertex}\nIn this section, we prove Conjecture \\ref{ABBH-conj} for vertex\ndecomposable graphs.\nWe first recall the definition of simplicial complex and vertex decomposable graph.\n\nA \\textit{simplicial complex} $\\Delta$ on $V = \\{x_1,\\ldots,x_n\\}$ is\na collection of subsets of $V$ such that: \n\\begin{enumerate}\n \\item $\\{x_i\\}\\in \\Delta $ for $i =1,\\ldots,n$, and\n \\item if $F \\in \\Delta$ and $G \\subseteq F$, then $G \\in \\Delta$.\n\\end{enumerate}\nElements of $\\Delta$ are called the \\textit{faces} of $\\Delta$, and the maximal elements, with \nrespect to inclusion, are called the facets. The link of a face $F$ in\n$\\Delta$ is $\\link_\\Delta(F) = \\{F' \\mid F' \\cup F \\text{ is a face in\n} \\Delta, ~F' \\cap F = \\emptyset \\}$.\n\n\nA simplicial complex $\\Delta$ is\nrecursively defined to be {\\em vertex decomposable} if it is either a\nsimplex or else has some vertex $v$ so that \n\\begin{enumerate}\n  \\item both $\\Delta \\setminus v$ and $\\link_\\Delta v$ are vertex decomposable, and\n  \\item no face of $\\link_\\Delta v$ is a facet of $\\Delta \\setminus v$.\n\\end{enumerate}\n\nThe \\textit{independence complex} of $G$, denoted $\\Delta(G)$, is the simplicial\ncomplex on $V(G)$ with face set \n$$\\Delta(G)=\\Big\\{F \\subseteq V(G) \\mid F \\text{ is an independent set of $G$ } \\Big\\}.$$\nA graph $G$ is said to be \\textit{vertex decomposable} if $\\Delta(G)$ is a\nvertex decomposable simplicial complex. \nIn \\cite{Wood2}, Woodroofe translated the notion of vertex decomposable\nfor graphs as follows.\n\\begin{definition} \\cite[Lemma 4]{Wood2}\n A graph $G$ is recursively defined to be vertex decomposable if \n $G$ is totally disconnected (with no edges) or if \n \\begin{enumerate}\n  \\item there is a vertex $x$ in $G$ such that $G \\setminus x$ and \n  $G \\setminus N_G[x]$ are both vertex decomposable, and\n  \\item no independent set in $G\\setminus N_G[x]$ is a maximal independent set \n  in $G\\setminus x$.\n \\end{enumerate}\n\\end{definition}\nA vertex $x$ which satisfies the second condition is called a \\textit{shedding vertex} of\n$G$.\n If $G$ is a vertex decomposable graph, then by \\cite[Theorem 2.5]{BFH15},\n $G \\setminus N_G[x]$ is a vertex decomposable graph, for any  $x \\in V(G)$.\n For any vertex decomposable graph $K$, set \n \\[\n\\s(K)=\\Big \\{ x \\in V(K) \\mid \\text{$x$ is a shedding vertex and \n$K \\setminus x$ is a vertex decomposable graph} \\Big\\}.\n\\]\n Note that if $K$ is vertex decomposable, then $\\s(K) \\neq \\emptyset$.\n \\begin{obs} \\label{reg-obs}\n Let $G$ be a vertex decomposable graph and $x \\in \\s(G)$. By \\cite[Theorem 4.2]{HaWood},\n $$\\reg(I(G))=\\max\\Big\\{\\reg(I(G \\setminus x)), ~\\reg(I(G \\setminus N_G[x]))+1\\Big\\}.$$\n Therefore, $\\reg(I(G \\setminus N_G[x]))+1 \\leq \\reg(I(G))$.\n \\end{obs}\n We prove Conjecture \\ref{ABBH-conj} for the class of vertex decomposable graphs.\n\\begin{theorem}\\label{upper-ver}\n  Let $G$ be a vertex decomposable graph. Then for all $q \\geq 1$,\n $$\\reg(I(G)^q) \\leq 2q+\\reg(I(G))-2.$$\n \\end{theorem}\n\\begin{proof}\nThis proof is also very similar to the proof of Theorem\n\\ref{weak-conj1}. We give a sketch of the proof here.  The proof is by\ninduction on $q$. If $q=1,$ then we are done.  Assume that $q >1$. By\napplying \\cite[Theorem 5.2]{banerjee} and using induction, it is\nenough to prove that for edges $e_1, \\ldots, e_s$ of $G$ (not\nnecessarily distinct), $\\reg((I(G)^{s+1}:e_1 \\cdots e_s)) \\leq\n\\reg(I(G))$ for all $s \\geq 1$.  We prove this by induction on $s$.\nLet $G'$ be the graph associated to $\\widetilde{(I(G)^{s+1}:e_1\\cdots\ne_s)}$, for any $e_1, \\ldots, e_s \\in E(G)$.\n\nLet $s = 1$. As in the proof of Theorem \\ref{weak-conj1}, we split the\nproof into two cases.\n\\vskip 1mm \\noindent\n\\textsc{Case 1:} Suppose $e_1\\cap \\s(G) \\neq \\emptyset$.\n\\vskip 1mm \\noindent \nThe proof is identical to \\textsc{Case 1} in Theorem \\ref{weak-conj1}.\nThe only difference in the proof is that\nwhile we used the hypothesis in Theorem \\ref{weak-conj1} to conclude\nthat $\\reg(J:y_i) < \\reg(I(G))$, here we use Observation \\ref{reg-obs}\nfor that conclusion.\n\\vskip 1mm \\noindent\n\\textsc{Case 2:} Suppose $e_1 \\cap \\s(G) = \\emptyset$. \n\\vskip 1mm \\noindent\nLet $|V(G)|=n$.  We proceed by induction on $n$. \nIf $G$ is a vertex decomposable with $n \\leq 4$, then\nit can be seen that $e \\cap \\s(G)\\neq \\emptyset$, for all $e \\in\nE(G)$. Therefore $n \\geq 5$.\n\n\\noindent\n \\begin{minipage}{\\linewidth}\n \n\\begin{minipage}{0.41\\linewidth} \n\\begin{figure}[H]\n \\begin{tikzpicture\n\\draw (3,3)-- (2,2);\n\\draw (3,3)-- (4,2);\n\\draw (2,2)-- (4,2);\n\\draw (5,3)-- (5,2);\n\\draw (5,2)-- (7,2);\n\\draw (7,2)-- (7,1);\n\\draw (5,2)-- (5,1);\n\\draw (5,1)-- (7,1);\n\\draw (4,2)-- (4,1);\n\\draw (2,2)-- (2,1);\n\\draw (2,1)-- (4,1);\n\\begin{scriptsize}\n\\fill [color=black] (3,3) circle (1.5pt);\n\\fill [color=black] (2,2) circle (2.5pt);\n\\fill [color=black] (4,2) circle (2.5pt);\n\\fill [color=black] (5,3) circle (1.5pt);\n\\fill [color=black] (5,2) circle (2.5pt);\n\\fill [color=black] (7,2) circle (1.5pt);\n\\fill [color=black] (7,1) circle (1.5pt);\n\\fill [color=black] (5,1) circle (1.5pt);\n\\fill [color=black] (4,1) circle (1.5pt);\n\\fill [color=black] (2,1) circle (1.5pt);\n\\end{scriptsize}\n\\end{tikzpicture}\n\\caption*{Vertex decomposable graphs with $|V(G)|=5$ and $\\bullet$\ndenote shedding vertices.}\\label{fig:vertexdecom}\n\\end{figure}\n\\end{minipage}\n\\begin{minipage}{0.59\\linewidth}\nSuppose $n=5$.\nIt can verified, manually or using computational packages such as\nMacaulay 2 \\cite{M2}, SimplicialDecomposability \\cite{Cook} or SAGE \\cite{sage}, that there are $20$ vertex decomposable\ngraphs (without isolated vertices) on $5$ vertices. Among these\ngraphs, all except two graphs satisfy the property that $G \\setminus\n\\s(G)$ is a set of isolated vertices. The two graphs for which $G\n\\setminus \\s(G)$ contain edges are given in the figure on the left.\nThen for any choice of $e \\in E(G \\setminus\n\\s(G)), ~ (I(G)^2 : e) = I(G)$ so that $G' = G$.\n\\end{minipage}\n\\end{minipage}\n\nNow assume that $n>5$. By induction, assume that if $H$ is a\nvertex decomposable graph with $|V(H)| < n$ and $e\n\\in E(H)$ such that $e \\cap S(H)= \\emptyset$, then $\\reg(I(H)^2 :\ne) \\leq \\nu(H)+1$. \nLet $x \\in \\s(G)$. By \\cite[Theorem 3.4]{Ha2}, \n\\[\\reg(I(G')) \\leq \\max\n\\Big\\{\\reg(I(G' \\setminus \\{x\\})), \\reg(I(G' \\setminus\nN_{G'}[x])+1\\Big\\}.\\]\nNow the proof is identical to that of the proof of \\textsc{Case 2} in\nTheorem \\ref{weak-conj1}. In this case also, we use Observation\n\\ref{reg-obs} to conclude $\\reg(I(G \\setminus N_G[x])) < \\reg(I(G))$.\n\nSuppose $s > 1$. Assume by induction that for any vertex decomposable graph $G$\nand edges $e_1,\\ldots, e_{s-1}$, $\\reg(I(G)^s : e_1 \\cdots e_{s-1}) \\leq \\reg(I(G))$.\nWe now prove that for edges $e_1, \\ldots, e_s$, $\\reg(I(G)^{s+1} :\ne_1\\cdots e_s) \\leq \\reg(I(G))$. Let $e_i = \\{a_i, b_i\\}$. If\n$\\deg_G(a_i) = 1$ or $\\deg_G(b_i) = 1$ for some $i$, then the\nassertion follows as in the proof of Theorem \\ref{weak-conj1}.\nTherefore, we may assume that $\\deg_G(a_i) \\geq 2$ and $\\deg_G(b_i)\n\\geq 2$ for all $1 \\leq i \\leq s$.\n\n  \\vskip 1mm \\noindent\n\\textsc{Case 3:} Suppose $e_i \\cap \\s(G) \\neq \\emptyset$, for some $1\n\\leq i \\leq s$. \n\\vskip 1mm \\noindent\nWithout loss of generality, assume that $e_s \\cap \\s(G) \\neq \\emptyset$ and $a_s \\in \\s(G)$.\nProceeding as in the proof Theorem \\ref{weak-conj1} following the same\nnotation, one gets \n\\begin{eqnarray*}\n\\reg(J,X_1,X_2)  &\\leq & \\reg(I(G' \\setminus N_{G'}[b_s]))\\leq \\reg(I(G \\setminus N_G[b_s])) \\leq\n  \\reg(I(G));  \\\\  \n  \\reg(J : y_i) & \\leq & \\reg(I((G \\setminus N_G[y_i,a_s])')) \\leq \\reg(I((G \\setminus N_G[a_s])')) \\\\\n  & \\leq&  \\reg(I(G \\setminus N_G[a_s])) < \\reg(I(G));\n  \\\\\n  \\reg(J,X_1 : z_i) &\\leq &\\reg(I((G \\setminus N_G[y_i,a_s])')) \\leq \\reg(I((G \\setminus N_G[a_s])'))\\\\\n  & \\leq & \\reg(I(G \\setminus N_G[a_s])) < \\reg(I(G)).\n\\end{eqnarray*}\nHere also, we use Observation \\ref{reg-obs} along with \nLemmas \\ref{tech_lemma},\n\\ref{ind-reg}, \\ref{even-lemma} and induction on $s$ for the above\nconclusions.\nUsing these inequalities, we conclude, as in the proof\nof Theorem \\ref{weak-conj1}, that $\\reg(J) \\leq \\reg(I(G))$.\n\n\\vskip 1mm \\noindent\n\\textsc{Case 4:}\nSuppose $e_i \\cap \\s(G) = \\emptyset$, for all $1 \\leq i \\leq s$.\n\nLet $x \\in \\s(G)$. By \\cite[Theorem 3.4]{Ha2}, \n\\[\\reg(I(G')) \\leq \\max\n\\Big\\{\\reg(I(G' \\setminus \\{x\\})), \\reg(I(G' \\setminus\nN_{G'}[x])+1\\Big\\}.\\]\nHere too, the proof is identical to \\textsc{Case 4} of Theorem\n\\ref{weak-conj1}.\n\nFinally, we obtain\n$\\reg(I(G)^{s+1} : e_1\\cdots e_s) \\leq \\reg(I(G))$. Therefore, for all\n$q \\geq 1$,\n$\\reg(I(G)^q) \\leq 2q + \\reg(I(G)) - 2$.\n\\end{proof}\n\nAs an immediate consequence of the above result, we obtain the linear\npolynomial corresponding to $\\reg(I(G)^q)$ for several \nclasses of graphs.\n\n\\begin{corollary}\\label{main-cor} If\n\\begin{enumerate}\n \\item $G$ is $C_5$-free vertex decomposable;\n \\item $G$ is chordal;\n \\item $G$ is sequentially Cohen-Macaulay bipartite,    ~~or\n \\item $G=H \\cup W(S)$, where $S \\subseteq V(H)$, $H \\setminus S$ is a\n   chordal graph and $H \\cup W(S)$ denotes the graph obtained from $H$ by adding a whisker to each vertex in $S$,\n\\end{enumerate}\nthen for all $q \\geq 1$, $$\\reg(I(G)^q)=2q+\\nu(G)-1.$$ \n\\end{corollary}\n\\begin{proof} (1) Follows from Theorem \\ref{upper-ver}, \\cite[Theorem 4.5]{selvi_ha} and\n\\cite[Lemma 2.3]{khosh_moradi}.\n\\vskip 1mm \\noindent\n(2) Since $G$ is chordal, it is $C_5$-free and by \\cite{Wood2} $G$ is\nvertex decomposable. By (1), $\\reg(I(G)^q)=2q+\\nu(G)-1$.\n \\vskip 1mm \\noindent\n (3) By \\cite[Theorem 2.10]{adam}, $G$ is vertex decomposable. \n  Since a bipartite graph is $C_5$-free, the assertion follows from\n  (1).\n  \\vskip 1mm \\noindent\n (4) First we show that $\\reg(I(G))=\\nu(G))+1$. By \\cite[Lemma 2.2]{Kat}, we have\n$\\nu(G))+1 \\leq \\reg(I(G)$. Therefore, it is enough to prove that $\\reg(I(G))\\leq \n\\nu(G)+1$. Set $|S|=m$.\n If $m=0$, then by \\cite[Corollary 6.9]{ha_adam}\n $\\reg(I(G))=\\nu(G)+1.$\n Suppose $m \\geq 1$. There is a vertex $x$ in $S$ such that $N_{G}(z_x)=\\{x\\}$. \n By \\cite[Theorem 3.4]{Ha2},\n $$\\reg(I(G)) \\leq \\max \\{\\reg(I(G \\setminus x)),\\reg(I(G \\setminus N_{G}[x]))+1\\}.$$\n By induction hypothesis on $m$, \n $\\reg(I(G \\setminus x)) \\leq \\nu(G)+1$ and $\\reg(I(G \\setminus N_{G}[x])))+1 \\leq \n \\nu(G \\setminus N_G[x])+2$. If $\\{f_1,\\ldots,f_t\\}$ is an induced matching of $G \\setminus N_G[x]$,\n then $\\{f_1,\\ldots,f_t,\\{x,z_x\\}\\}$ is an induced matching of $G$.\n Therefore $\\nu(G \\setminus N_G[x])+1 \\leq \\nu(G)$. Hence\n $ \\reg(I(G))=\\nu(G)+1.$\n By \\cite[Corollary 4.6]{BFH15}, $G$ is a vertex decomposable graph. Therefore, by\n Theorem \\ref{upper-ver} and \\cite[Theorem 4.5]{selvi_ha},\n $\\reg(I(G)^q)=2q+\\nu(G)-1.$\n\\end{proof}\n\n\n\\vskip 2mm \\noindent\n\\textbf{Acknowledgement:} We would like to thank  Fahimeh Khosh-Ahang,\nHuy T{\\`a}i H\\`a, Adam Van Tuyl and Russ Woodroofe for several\nclarifications on our doubts on their results. We also would like to\nthank Arindam Banerjee and Huy T\\`ai H\\`a for pointing out an error in\none of the proofs in an earlier version of the manuscript. We\nextensively used \\textsc{Macaulay2}, \\cite{M2}, \\textsc{SAGE},\n\\cite{sage} and the package \\textsc{SimplicialDecomposability},\n\\cite{Cook}, for testing our our computations.\n\n\n\n\n\\bibliographystyle{abbrv} \n", "meta": {"timestamp": "2018-05-04T02:12:24", "yymm": "1805", "arxiv_id": "1805.01412", "language": "en", "url": "https://arxiv.org/abs/1805.01412", "abstract": "Let $G$ be a finite simple graph and $I(G)$ denote the corresponding edge ideal. In this paper, we obtain upper bounds for the Castelnuovo-Mumford regularity of $I(G)^q$ in terms of certain combinatorial invariants associated with $G$. We also prove a weaker version of a conjecture by Alilooee, Banerjee, Beyarslan and Hà on an upper bound for the regularity of $I(G)^q$ and we prove the conjectured upper bound for the class of vertex decomposable graphs. Using these results, we explicitly compute the regularity of $I(G)^q$ for several classes of graphs.", "subjects": "Commutative Algebra (math.AC); Combinatorics (math.CO)", "title": "Upper bounds for the regularity of powers of edge ideals of graphs", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9873750477464142, "lm_q2_score": 0.8539127492339909, "lm_q1q2_score": 0.8431321415461835}}
{"url": "https://arxiv.org/abs/1805.10510", "title": "A modification of the Chang-Wilson-Wolff Inequality via the Bellman Function", "abstract": "We describe the Bellman function technique for proving sharp inequalities in harmonic analysis. To provide an example along with historical context, we present how it was originally used by Donald Burkholder to prove $L^p$ boundedness of the $\\pm 1$ martingale transform. Finally, with Burkholder's result as a blueprint, we use the Bellman function to prove a new result related to the Chang-Wilson-Wolff Inequality.", "text": "\\section{Introduction}\n\nThe Bellman function technique, named for applied mathematician Richard Bellman, is a tool that has been imported from the applied field of stochastic optimal control, and is now being used to tackle problems in probability and harmonic analysis. It was introduced to the world of analysis by Donald Burkholder, who in \\cite{Burkholder} used it to prove that the $\\pm$1 transform of a martingale is a bounded operator on $L^p$. We will borrow Burkholder's ideas to prove a new result concerning the exponential integrability of dyadic martingales.\n\nSection 2 is expository. We briefly summarize Burkholder's use of the Bellman function to prove a sharp martingale inequality. This will serve as homage to Burkholder, the pioneer, and provide some historical context. As importantly, it will give a template upon which we will build the proof of our main result. The Bellman function technique is much easier to demonstrate than it is to describe abstractly.\n\nIn section 3, we introduce a well known inequality from harmonic analysis due to Chang, Wilson, and Wolff which classifies the order of local integrability of a function whose dyadic square function is bounded. Their result says that given a function $f:[0,1) \\to \\mathbb{R}$ and its dyadic square function $Sf$,\n\\begin{equation*}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\leq e^{\\frac{1}{2}\\|(Sf)^2(x)\\|_{L^\\infty}}\n\\end{equation*}\nNext, we address a related question from \\cite{SV}. Namely, we explore whether there exists a constant $\\alpha$ such that\n\\begin{equation*}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\leq \\int_0^1 e^{\\alpha(Sf)^2(x)}dx\n\\end{equation*}\nand if so, what the smallest valid choice of $\\alpha$ is. Stated probabilistically, given a dyadic martingale with $f_0 = 0$, what is the smallest $\\alpha$ such that\n\\begin{equation*}\n\\mathbb{E}e^{f_n} \\leq \\mathbb{E}e^{\\alpha(Sf_n)^2}\n\\end{equation*}\nWe use the Bellman function to prove, without constructing an explicit example, that $\\alpha = 2$ makes this inequality sharp.\n\nLastly, in section 4, we provide an alternate proof of our inequality by applying Cauchy-Schwarz to a result known as Rubin's lemma. We attempt to construct an example of a martingale which shows $\\alpha = 2$ is sharp, but come up short. Our example pushes $\\alpha$ up to $\\log_2(e) \\approx 1.44$, but finding the extremal martingale is left for future work.\n\n\\section{Bellman Function Technique}\n\nWe begin by illustrating the utility of the Bellman function. We shall summarize Burkholders arguments, which we will repurpose for our own problem later on. The exposition roughly follows that of \\cite{Osekowski}.\n\n\\begin{definition}\nLet $(\\Omega, \\mathcal{F}, \\mathbb{P})$ be a probability space filtered by $\\{\\mathcal{F}_n\\}$. Then the sequence of random variables $\\{f_n\\}$ is called a martingale if for each $n \\in \\mathbb{N}$\n\\begin{enumerate}\n\\item $f_n$ is $\\mathcal{F}_n$ measurable\\\\\n(i.e., $\\{f_n\\}$ is adpapted to the filtration $\\{\\mathcal{F}_n\\}$)\n\\item $\\|f_n\\|_{L^1} < \\infty$\n\\item $E[f_{n+1}|\\mathcal{F}_n] = f_n$\n\\end{enumerate}\n\\end{definition}\n\nIf we replace the equality in condition 3 with $\\leq$ or $\\geq$, $\\{f_n\\}$ is called a supermartingale or submartingale respectively.\n\n\\begin{definition}\nLet $\\{f_n\\}$ be a martingale. Define $df_0 = f_0$ and $df_n = f_n - f_{n-1}$ for $n>0$. $\\{df_n\\}$ is called the difference sequence of ${f_n}$.\n\\end{definition}\n\nNote that $\\{df_n\\}$ is adapted to $\\mathcal{F}_n$ and $E[df_{n+1}|\\mathcal{F}_n] = 0$. It is often useful to express a martingale as the sum of its difference sequence, $f_n = \\sum_{k=0}^n df_k$.\n\n\\begin{definition}\nLet $\\{f_n\\}$ be a martingale, and define\n$$g_n = \\sum_{k=0}^n \\epsilon_k df_k$$\nwhere $\\{\\epsilon_k\\}$ is a deterministic sequence, all of whose terms are $\\pm1$. $\\{g_n\\}$ is called a $\\pm 1$ transform of $\\{f_n\\}$.\n\\end{definition}\n\nIn \\cite{Burkholder}, Burkholder used the Bellman function technique to prove the following result.\n\n\\begin{theorem}\nThere is a constant $\\beta_p$ such that if $\\{g_n\\}$ is a $\\pm 1$ transform of $\\{f_n\\}$, then for all $n \\in \\mathbb{N}$ and $0<p<\\infty$\n\\begin{equation}\n\\|g_n\\|_p \\leq \\beta_p \\|f_n\\|_p\n\\end{equation}\n\\end{theorem}\n\nIn other words, each $\\pm 1$ transform is a bounded operator on $L^p$ for $0<p<\\infty$. Theorem 1 has an interesting corollary, which highlights the connection between harmonic analysis and probability. Theorem 1 implies that the Haar basis for $L^p([0,1))$ is unconditional since the partial sums of a Haar series are a martingale on $([0,1),\\mathcal{B},|\\cdot|)$ (see \\cite{Burkholder} or \\cite{Osekowski} for details). Thinking about functions as random variables is sometimes a useful perspective in analysis as it allows us to import tools from probability where it's not obvious that they belong.\n\nWe now outline Burkholder's proof of Theorem 1 in order to demonstrate the Bellman function technique and provide a blueprint. Its essence is to relate the validity of an inequality to the existence of a special function. Often the existence of the function is easier to prove or disprove than the given inequality. We will henceforth assume that $\\{f_n\\}$ is a simple martingale (each $f_n$ takes on finitely many values and eventually $f_n = f_{n+1} = \\dots$). Passage to the general case follows from an approximation argument.\n\n\\begin{theorem}\nSuppose there exists a function $B:\\mathbb{R}^2 \\to \\mathbb{R}$ with the following three properties.\n\\begin{enumerate}\n\\item (Majorization) $B(x,y) \\geq x^p - \\beta_p^p y^p \\coloneqq V(x,y)$\n\\item (Concavity) For all $x,y,t_1,t_2 \\in \\mathbb{R}, \\epsilon = \\pm 1$, and $\\alpha \\in (0,1)$ such that $\\alpha t_1 + (1 - \\alpha) t_2 = 0$, we have\n$$\\alpha B(x + t_1,y + \\epsilon t_1) + (1-\\alpha)B(x + t_2, y + \\epsilon t_2) \\leq B(x,y)$$\n\\item (Initial condition) $B(x,\\pm x) \\leq 0$\n\\end{enumerate}\nThen (2.1) holds.\n\\end{theorem}\n\nAs promised, we've reduced the veracity of (2.1) to the existence of a certain function with special properties. Note that, due to the majorization property, the existence of $B$ depends on our choice of $\\beta$ and $p$. This is to be expected because (1) may hold for some $\\beta$ and $p$ but not others.\n\n$B$ is actually not the Bellman function, but a pointwise majorant of it. The definition of Bellman function and its relationship to our $B$ will be discussed momentarily.\n\nProperty 1 states that $B$ dominates a function $V$ whose definition is suggested by the inequality we're after. Note that (2.1) is equivalent to $\\mathbb{E}V(f_n,g_n) \\leq 0$. Property 2 says that $B$ is \"diagonally concave,\" i.e., it is concave along the lines of slope $\\pm 1$. This implies $\\mathbb{E}B(x + \\xi, y \\pm \\xi) \\leq B(x,y)$ for all mean zero random variables $\\xi$ by Jensen's inequality. The form of this concavity condition also varies with the inequality to be proven. It is chosen so that $\\{B(f_n,g_n)\\}$ is a supermartingale.\n\n\\begin{lemma}\nIf $B$ satisfies condition 2 in Theorem 2.5, then $\\{B(f_n,g_n)\\}$ is a supermartingale.\n\\end{lemma}\n\n\\begin{proof}\n\\begin{align}\n\\mathbb{E}[B(f_{n+1},g_{n+1})|\\mathcal{F}_n] &= \\mathbb{E}[B(f_n + df_{n+1},g_n \\pm df_{n+1})|\\mathcal{F}_n]\\\\\n&\\leq \\mathbb{E}[B(f_{n},g_{n})|\\mathcal{F}_n]\\\\\n&= B(f_n,g_n)\n\\end{align}\nwhere (2.2) uses that $g_n$ is a $\\pm 1$ transform of $f_n$ and (2.3) is from applying condition 2 in Theorem 2.5 conditionally. (2.4) is because $(f_n,g_n)$ is $\\mathcal{F}_n$ measurable.\n\\end{proof}\n\nWe now prove Theorem 2.5.\n\n\\begin{proof}\nRecall that it suffices to show $\\mathbb{E}V(f_n,g_n) \\leq 0$.\n\\begin{align}\n\\mathbb{E}V(f_n,g_n) &\\leq \\mathbb{E}B(f_n,g_n)\\\\\n&=\\mathbb{E}\\big[\\mathbb{E}[B(f_n,g_n)|\\mathcal{F}_{n-1}]\\big]\\\\\n&\\leq \\mathbb{E}B(f_{n-1},g_{n-1})\n\\end{align}\nwhere (2.5) follows from majorization (2.6) is because conditional expectation preserves expectation. (2.7) is from the fact that $\\{B(f_n,g_n)\\}$ is a supermartingale.\nRepeating the argument n times, we get\n$$\\mathbb{E}V(f_n,g_n) \\leq \\mathbb{E}B(f_0,g_0) = \\mathbb{E}B(df_0,\\pm df_0) \\leq 0$$\nThe final inequality is from the initial condition.\n\\end{proof}\n\nWith Theorem 2.5 proved, we can now prove Theorem 2.4 by producing an appropriate $B$. However, we can actually do more. In fact, given a $0<p<\\infty$ and $\\beta_p$, if no such $B$ exists, then (2.1) is false.\n\n\\begin{theorem}\nGiven $0<p<\\infty$ and $\\beta_p$, (2.1) holds if and only if there exists a function $B$ with the three properties from Theorem 2.\n\\end{theorem}\n\n\\begin{proof}\nWe know that if $B$ exists, then (2.1) holds by Theorem 2.5. Now we must show that if (2.1) holds, then $B$ exists.\n\nLet $\\mathcal{M}(x,y)$ be the set of all $\\mathbb{R}^2$ valued martingales $(f_n,g_n)$ such that $(f_0,g_0) \\equiv (x,y)$ and $dg_n = \\pm df_n$ for $n \\geq 1$. We define the Bellman function\n\n$$\\mathcal{B}(x,y) \\coloneqq \\sup\\{\\mathbb{E}V(f_n,g_n) : (f_n,g_n)\\in M(x,y)\\}$$\n\nWe will show that $\\mathcal{B}$ is the desired function $B$ possessing the three properties. Majorization is straightforward. Observe that the deterministic pair $(x,y) \\in \\mathcal{M}(x,y)$. The initial condition $\\mathcal{B}(x,\\pm x) \\leq 0$ follows from (2.1). To show concavity, we use a \"splicing argument.\" Take $x,y,t_1,t_2,\\alpha,\\epsilon$ as in the statement of the concavity condition. Choose any $(f_n^a,g_n^a) \\in \\mathcal{M}(x + t_1, y + \\epsilon t_1)$ and $(f_n^b,g_n^b) \\in \\mathcal{M}(x + t_2, y + \\epsilon t_2)$. We may assume the pairs are given on $([0,1),\\mathcal{B},|\\cdot|)$. We will define another martingale by \"splicing\" these two. Let $(f_0,g_0) \\equiv (x,y)$ and for $n \\geq 0$\n\\[ (f_{n+1},g_{n+1})(\\omega) = \n\t\\begin{cases} \n\t\t(f_n^a,g_n^a)(\\frac{\\omega}{\\alpha}) & \\omega \\in [0,\\alpha) \\\\\n\t\t(f_n^b,g_n^b)(\\frac{\\omega - \\alpha}{1 - \\alpha}) & \\omega \\in [\\alpha,1) \\\\\n\t\\end{cases}\n\\]\nOne can check that $\\{(f_n,g_n)\\} \\in \\mathcal{M}(x,y)$, and\n\\begin{align*}\n\\mathcal{B}(x,y) \\geq \\mathbb{E}V(f_n,g_n) = \\alpha\\mathbb{E}V(f_n^a,g_n^a) + (1 - \\alpha)\\mathbb{E}V(f_n^b,g_n^b)\n\\end{align*}\nTaking the supremum over all such $(f_n^a,g_n^a)$ and $(f_n^b,g_n^b)$ yields\n\\begin{align*}\n\\mathcal{B}(x,y) \\geq \\alpha \\mathcal{B}(x + t_1,y + \\epsilon t_1) + (1-\\alpha)\\mathcal{B}(x + t_2, y + \\epsilon t_2)\n\\end{align*}\n\nThe last thing that must be checked is that $\\mathcal{B}$ is finite on all of $\\mathbb{R}^2$. We know $\\mathcal{B} \\geq V > -\\infty$ so we only need to check $\\mathcal{B}(x,y) < \\infty$. The initial condition says that $\\mathcal{B}(x,\\pm x) \\leq 0$. Now suppose $|x|\\neq |y|$ and let $(f_n,g_n)\\in \\mathcal{M}(x,y)$. We construct another martingale $(f_n',g_n')$ as follows (again we may assume our martingales are defined on $([0,1),\\mathcal{B},|\\cdot|)$.\n\n\\begin{align*}\n(f'_0,g'_0) &\\equiv (\\frac{x+y}{2},\\frac{x+y}{2}) \\\\\n(f'_1,g'_1)(\\omega) &=\n\t\\begin{cases}\n\t\t(x,y) & \\omega\\in [0,\\frac{1}{2}) \\\\\n\t\t(y,x) & \\omega\\in [\\frac{1}{2},1) \\\\\n\t\\end{cases} \\\\\n(f'_n,g'_n)(\\omega) &=\n\t\\begin{cases}\n\t\t(f_{n-1},g_{n-1})(2\\omega) & \\omega\\in [0,\\frac{1}{2}) \\\\\n\t\t(y,x) & \\omega\\in [\\frac{1}{2},1)\\\\\n\t\\end{cases}\n\\end{align*}\n\\noindent\nwhere the last equality holds for $n \\geq 2$. Note that $f'$ is a $\\pm 1$ transform of $g'$ and \n$$0 \\geq \\mathbb{E}V(f_n',g_n') = \\frac{1}{2}V(y,x) + \\frac{1}{2}\\mathbb{E}V(f_{n-1},g_{n-1})$$\n\\noindent\nTaking the supremum over all $(f_n,g_n)\\in \\mathcal{M}(x,y)$ gives $\\mathcal{B}(x,y) \\leq -V(y,x) < \\infty$ and we're done.\n\n\\end{proof}\n\nEquipped with Theorem 2.7, we can do more than just prove our inequality. We can find the optimal constant $\\beta_p$. If $B$ does not exist when $\\beta_p < C_p$ but $B$ does exist when $\\beta_p \\geq C_p$, then $C_p$ is optimal. We will not construct $B$ here because our intention is not to reproduce Burkholder's result, but rather demonstrate the utility of his approach and provide a blueprint for proving our new inequality.\n\n\n\\section{A modification of the Chang-Wilson-Wolff Inequality}\n\n\nLet $f \\in L^1(I)$ for some real interval $I$. Let $\\mathcal{F}_n$ be the $\\sigma$ algebra generated by the dyadic subintervals of $I$ of length $|I|2^{-n}$ and let $\\mathcal{F} = \\bigcup \\mathcal{F}_n$. Then $f_n = \\mathbb{E}[f|\\mathcal{F}_n]$ is a martingale on the probability space $(I,\\mathcal{F},\\frac{|\\cdot|}{|I|})$. Due to the dyadic filtration, such a martingale is called a dyadic martingale.\n\nNow define $Sf_n = \\|\\{df_n\\}\\|_{\\ell^2} = \\sqrt{\\sum_{k=1}^n (df_k)^2}$ and let $Sf = \\lim\\limits_{n \\to \\infty} Sf_n$. In the context of probability, the sequence $\\{Sf_n\\}$ is called the quadratic variation of the martingale $\\{f_n\\}$. In the Littlewood-Paley theory of harmonic analysis, $Sf$ is called the dyadic square function of $f$ and is the discrete counterpart of the Lusin area function, whose definition would take us too far from our present goal. For a description of the connection between the square function and the Luzin area function, refer to \\cite{Llorente}. For a survey of the role of square functions in harmonic analysis, consult \\cite{Stein}.\n\nIn \\cite{CWW}, the authors answer a question posed by the illustrious harmonic analyst Elias Stein. The question concerns the size of the functions in a certain subspace of BMO: What is the sharp order of local integrability of a function $f$ with a pointwise bounded square function $Sf$? Suppose for simplicity and concreteness that $f \\in L^1([0,1))$. The authors show that\n\\begin{equation}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\leq e^{\\frac{1}{2}\\|Sf\\|^2_{L_{[0,1)}^\\infty}}\n\\end{equation}\nThis result, called the Chang-Wilson-Wolff inequality, can be used to show that if $Sf$ is bounded, then $f$ is exponentially square integrable \\cite{Pipher}.\n\nIn \\cite{SV}, the authors explore a related question: Is there a constant $\\alpha$ such that \n\n\\begin{equation}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\leq \\int_0^1 e^{\\alpha(Sf)^2(x)}dx\n\\end{equation}\nInspired by their work, our goal in this section is to use the Bellman function technique to find the smallest such $\\alpha$. In the language of probability, given a dyadic martingale with $f_0 = 0$, is there a constant $\\alpha$ such that for each $n \\in \\mathbb{N}$\n\\begin{equation}\n\\mathbb{E}e^{f_n} \\leq \\mathbb{E}e^{\\alpha(Sf_n)^2}\n\\end{equation}\nIf so, what is the smallest such $\\alpha$? We shall use Burkholder's approach as a blueprint. As before, the veracity of (3.3) can be recast as a question of the existence of a special function satisfying three properies. The nature of these properties are gleaned from the details of the inequality we are after.\n\n\\begin{theorem}\n(3.3) holds if and only if there exists a function $B:\\mathbb{R} \\times [0,\\infty) \\to \\mathbb{R}$ with the following three properties.\n\\begin{enumerate}\n\\item (Majorization) $B(x,y) \\geq e^x-e^{\\alpha y} \\coloneqq V(x,y)$\n\\item (Concavity) $\\frac{B(x + \\delta, y + \\delta^2) + B(x - \\delta, y + \\delta^2)}{2} \\leq B(x,y)$ for any $\\delta \\in \\mathbb{R}$\n\\item (Initial condition) $B(0,0) \\leq 0$\n\\end{enumerate}\n\\end{theorem}\n\n\\noindent\n\\textbf{Remark:} As before, property 2 has a probabilistic interpretation. Given any discrete random variable $\\xi$ with $\\mathbb{P}(\\xi = \\delta) = \\mathbb{P}(\\xi = -\\delta) = \\frac{1}{2}$, we have $\\mathbb{E}B(x + \\xi, y + \\xi^2) \\leq B(x,y)$.\n\n\\begin{proof}\nThe proof is the same as those of Theorems 2.5 and 2.7 in the previous section, only adapted to the present inequality. Keep in mind that we are assuming throughout $\\{f_n\\}$ is a simple dyadic martingale with $f_0=0$. First, assume $B$ exists. Then $B(f_n,(Sf)^2_n)$ is a supermartingale. Indeed,\n\\begin{align}\n\\mathbb{E}[B(f_{n+1},(Sf)^2_{n+1})|\\mathcal{F}_n] &= \\mathbb{E}[B(f_n + df_{n+1}, (Sf)^2_n + (df)^2_{n+1})|\\mathcal{F}_n]\\\\\n&\\leq \\mathbb{E}[B(f_{n},g_{n})|\\mathcal{F}_n]\\\\\n&= B(f_n,g_n)\n\\end{align}\nwhere (3.5) is from applying the concavity property. Therefore, we have\n\\begin{align}\n\\mathbb{E}V(f_n,(Sf)^2_n) &\\leq \\mathbb{E}B(f_{n},(Sf)^2_{n})\\\\\n&= \\mathbb{E}\\big[\\mathbb{E}[B(f_{n},(Sf)^2_{n})|\\mathcal{F}_{n-1}]\\big]\\\\\n&\\leq \\mathbb{E}B(f_{n-1},(Sf)^2_{n-1})\n\\end{align}\nwhere (3.7) follows from majorization and (3.9) from the fact that $\\{B(f_n,(Sf)^2_n)\\}$ is a supermartingale.\nRepeating the argument n times, we get\n$$\\mathbb{E}V(f_n,(Sf)^2_n) \\leq \\mathbb{E}B(f_0,(Sf)^2_0) = \\mathbb{E}B(0,0) \\leq 0$$\nTo prove the other direction, we assume (3.3) holds, and construct $B$. As before, we define the Bellman function $\\mathcal{B}(x,y) \\coloneqq \\sup\\{\\mathbb{E}V(f_n,(Sf)^2_n) : (f_n,(Sf)^2_n)\\in \\mathcal{M}(x,y)\\}$ where $\\mathcal{M}(x,y)$ is the set of all $\\mathbb{R} \\times [0,\\infty)$ valued processes $(f_n,(Sf)^2_n)$ such that $\\{f_n\\}$ is a dyadic martingale with $f_0 = x$ and $(Sf)^2_n = y + \\sum_{k=1}^n (df_k)^2$ where $df_0 \\coloneqq 0$. Again, as before, we show that $\\mathcal{B}$ satisfies the three properties.\n\nMajorization follows from observing the constant pair $(x,y) \\in \\mathcal{M}(x,y)$ and the initial condition follows from (3.3). Once again, we can get concavity by a splicing argument.\n\nChoose any $(f_n^a,(Sf_n^a)^2) \\in \\mathcal{M}(x + \\delta, y + \\delta^2)$ and $(f_n^b,(Sf_n^b)^2) \\in \\mathcal{M}(x - \\delta, y + \\delta^2)$. We may assume the pairs are given on $([0,1),\\mathcal{B},|\\cdot|)$. We will define another martingale by \"splicing\" these two. Let $(f_0,(Sf_0)^2) \\equiv (x,y)$ and for $n \\geq 0$\n\\[ (f_{n+1},(Sf_{n+1})^2)(\\omega) = \n\t\\begin{cases} \n\t\t(f_n^a,(Sf_n^a)^2)(2\\omega) & \\omega \\in [0,\\frac{1}{2}) \\\\\n\t\t(f_n^b,(Sf_n^b)^2)(2(\\omega - \\frac{1}{2})) & \\omega \\in [\\frac{1}{2},1) \\\\\n\t\\end{cases}\n\\]\nOne can check that $\\{(f_n,(Sf_n)^2)\\} \\in \\mathcal{M}(x,y)$, and\n\\begin{align*}\n\\mathcal{B}(x,y) \\geq \\mathbb{E}V(f_n,(Sf_n)^2) = \\frac{\\mathbb{E}V(f_n^a,(Sf_n^a)^2) + \\mathbb{E}V(f_n^b,(Sf_n^b)^2)}{2}\n\\end{align*}\nTaking the supremum over all such $(f_n^a,(Sf_n^a)^2)$ and $(f_n^b,(Sf_n^b)^2)$ yields\n\\begin{align*}\n\\frac{\\mathcal{B}(x + \\delta, y + \\delta^2) + \\mathcal{B}(x - \\delta, y + \\delta^2)}{2} \\leq \\mathcal{B}(x,y)\n\\end{align*}\n\\noindent\nFinally, we must show $\\mathcal{B}$ is finite on $\\mathbb{R}\\times [0,\\infty)$. We know $-\\infty < V \\leq \\mathcal{B}$, so we only need to show $\\mathcal{B} < \\infty$. It follows from the definition of $\\mathcal{B}$ that $\\mathcal{B}(x + \\delta, y + \\frac{\\delta}{\\alpha}) = e^\\delta \\mathcal{B}(x,y)$ for any $\\delta \\in \\mathbb{R}$ such that $y + \\frac{\\delta}{\\alpha} \\geq 0$. Hence it suffices to show that $\\mathcal{B}$ is finite along the x axis. Using the concavity property, we have\n\n\\begin{align*}\n\\mathcal{B}(x,0) &\\geq \\frac{\\mathcal{B}(x + \\delta, \\delta^2) + \\mathcal{B}(x - \\delta, \\delta^2)}{2}\\\\\n&= \\frac{e^{\\alpha\\delta^2}\\mathcal{B}(x + \\delta - \\alpha\\delta^2, 0) + e^{\\alpha\\delta^2}\\mathcal{B}(x - \\delta - \\alpha\\delta^2, 0)}{2}\n\\end{align*}\n\\noindent\nTherefore,\n\\begin{align*}\n\\mathcal{B}(x+\\delta-\\alpha\\delta^2, 0) &\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}(x,0) - \\mathcal{B}(x-\\delta-\\alpha\\delta^2, 0)\\\\\n&\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}(x,0) - V(x-\\delta-\\alpha\\delta^2, 0)\n\\end{align*}\n\\noindent\nThe initial condition says that $\\mathcal{B}(0,0) \\leq 0,$ and we can see from the definition that $\\mathcal{B}$ is non-decreasing in the positive x direction, so $\\mathcal{B}(x,0) \\leq 0$ for all $x \\leq 0$. Furthermore, this monotonicity means it suffices to show $\\mathcal{B}(x_n,0) < \\infty$ for some sequence $x_n \\to \\infty$. If we take $\\delta$ small enough that $\\delta - \\alpha\\delta^2 > 0,$ we have \n\\begin{align*}\n\\mathcal{B}(\\delta-\\alpha\\delta^2, 0) &\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}(0,0) - V(-\\delta-\\alpha\\delta^2, 0) < \\infty \\\\\n\\mathcal{B}(2(\\delta-\\alpha\\delta^2), 0) &\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}(\\delta-\\alpha\\delta^2,0) - V(-2\\alpha\\delta^2, 0) < \\infty\\\\\n\\mathcal{B}(3(\\delta-\\alpha\\delta^2), 0) &\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}(2(\\delta-\\alpha\\delta^2),0) - V(\\delta - 3\\alpha\\delta^2, 0) < \\infty\\\\\n\\dots\\\\\n\\mathcal{B}(n(\\delta-\\alpha\\delta^2), 0) &\\leq 2e^{-\\alpha\\delta^2}\\mathcal{B}((n-1)(\\delta-\\alpha\\delta^2),0) - V((n-2)\\delta - n\\alpha\\delta^2, 0) < \\infty\\\\\n\\end{align*}\n\n\\end{proof}\n\n\\subsection*{Searching for a Bellman Function Candidate}\n\nWhat remains is to find a function $B$ so that we may apply Theorem 3.1. Our reasoning along the way needn't be rigorous or even correct because once we arrive at a candidate $B$, we will prove that it has the three properties. The purpose of this section is to demonstrate how one might approach the task of searching for $B$. Our approach will be to build a PDE. It should be noted that, a priori, we don't even know that a differentiable $B$ exists, but the purpose of this section is not to rigorously prove anything. Rather, it's to explain how one might arrive at a candidate, which we can then test for the three required properties.\n\nThe concavity condition $\\frac{B(x - \\delta, y + \\delta^2) + B(x + \\delta, y + \\delta^2)}{2} \\leq B(x,y)$ suggests that $B$ is concave along parabolic paths. Thus, we expect that $\\frac{d^2}{d\\delta^2}B(x + \\delta, y + \\delta^2) \\leq 0$.\n\\begin{align*}\n\\frac{d^2}{d\\delta^2}B(x + \\delta, y + \\delta^2) &= B_{xx}(x+\\delta, y + \\delta^2) + B_{xy}(x+\\delta, y + \\delta^2)4\\delta\\\\\n&+ B_{yy}(x+\\delta, y + \\delta^2)2\\delta^2 + 2B_y(x+\\delta, y + \\delta^2)\\\\\n&\\leq 0\n\\end{align*}\nEvaluating at $\\delta = 0$ yields\n\\begin{equation}\nB_{xx}(x,y) + 2B_y(x,y) \\leq 0\n\\end{equation}\nSo $B$ is a subsolution of the reverse heat equation.\n\nA common strategy is to search for Bellman candidates which share certain homogeneity properties with $V$. In our case, we have\n$$V\\big(x + \\delta,y + \\frac{\\delta}{\\alpha}\\big) = e^{x+\\delta} - e^{\\alpha(y + \\frac{\\delta}{\\alpha})} = e^\\delta V(x,y)$$\nLet $\\delta = -\\alpha y$. Then\n$$V(x - \\alpha y, 0) = e^{-\\alpha y}V(x,y)$$\nNote that functions with this property are completely determined by their values along the x axis. We shall search for a $B$ that shares this property.\\\\\nDefine $f(x) = B(x, 0)$ and assume $B$ has the form\n\\begin{equation}\nB(x,y) = e^{\\alpha y}f(x-\\alpha y)\n\\end{equation}\nPlugging (3.11) into (3.10) and replacing the inequality with an equation gives\n\\begin{equation}\ne^{\\alpha y}f''(x-\\alpha y) - 2\\alpha e^{\\alpha y} f'(x-\\alpha y) +2\\alpha e^{\\alpha y} f(x-\\alpha y) = 0\n\\end{equation}\nDividing through by $e^{\\alpha y}$ reveals a second order ODE with constant coefficients. The solutions to the characteristic equation are $\\alpha \\pm \\sqrt{\\alpha^2 -2\\alpha}$. If $\\alpha < 2$ then these solutions are complex, meaning $f$ oscillates. Thus $B(x,0)=f(x)$ will have no hope of dominating $V(x,0) = e^x - 1$. On the other hand, if $\\alpha = 2$, then $f(x) = C_1xe^{2x} + C_2e^{2x}$. We must select $C_1$  and $C_2$ such that $f(x) = B(x,0) \\geq V(x,0) = e^x - 1$ (majorization) and $f(0) = B(0,0) \\leq V(0,0) = 0$ (initial condition). Given these constraints, we find $f(x)=xe^{2x}$ and so our Bellman function candidate is $B(x,y) = (x - 2y)e^{2x - 2y}$. Unfortunately, this is not a valid $B$ as the concavity property is violated. For example, $\\frac{B(1 - \\delta, 1 + \\delta^2) + B(1 + \\delta, 1 + \\delta^2)}{2} \\to 0$ as $\\delta \\to \\infty$ but $B(1,1) = -1$. Interestingly, although this line of thinking produced the wrong $B$, it produced the correct $\\alpha$. It turns out $2$ is the sharp constant in (3.3).\n\nOne problem is that, assuming $B$ is smooth, (3.10) is a necessary but insufficient condition for concavity along parabolic paths. It only tests for concavity at the vertex of each path. Also, turning our differential inequality\n\\begin{equation}\nf'' - 2\\alpha f' +2\\alpha f \\leq 0\n\\end{equation}\ninto an equation was unjustified even though doing so produced the optimal $\\alpha$.\n\nAt this point, there is no shame in resorting to guess and check. That is, looking for solutions to (3.13) with $f(0) = 0$ and $f'(0)=1$ (to ensure compliance with the majorization property and the initial condition) and testing the corresponding $B(x,y)$ for the concavity property. $f(x) = e^{2x} - e^x$ is such a function (for $\\alpha = 2$) and the associated Bellman candidate is $B(x,y) = e^{2x - 2y} - e^x$. If we prove this $B$ has the three desired properties, we shown (3.3) holds with $\\alpha = 2$.\n\n\n\\begin{theorem}\nSuppose $\\{f_n\\}$ is a dyadic martingale with $f_0=0$. Then\n$$\\mathbb{E}e^{f_n} \\leq \\mathbb{E}e^{2(Sf_n)^2}$$\n\\end{theorem}\n\n\\begin{proof}\n$B(x,y) = e^{2x - 2y} - e^x$ obeys the three properties from Theorem 3.1. We begin with majorization. Using the inequality $1 - e^t \\leq e^{-t} - 1$ we have\n\\begin{align*}\nV(x,y) = e^x - e^{2y} = e^x(1 - e^{2y - x}) \\leq e^x(e^{x - 2y} - 1) = B(x,y)\n\\end{align*}\nNext, we show that $B$ satisfies the concavity condition.\n\\begin{align*}\n&\\frac{B(x+\\delta,y+\\delta^2)+B(x-\\delta,y+\\delta^2)}{2}\\\\\n&= \\frac{e^{2(x+\\delta) - 2(y+\\delta^2)} - e^{x+\\delta} + e^{2(x-\\delta) - 2(y+\\delta^2)} - e^{x-\\delta}}{2}\\\\\n&= e^{2x - 2y - 2\\delta^2}\\cosh2\\delta - e^x\\cosh\\delta\\\\\n&\\leq e^{2x - 2y} - e^x = B(x,y)\n\\end{align*}\nIn the last line, we use the inequality $1 \\leq \\cosh t \\leq e^\\frac{t^2}{2}$.\nLastly, the initial condition is immediate. $B(0,0) = 0$.\n\\end{proof}\n\nAt this point, we've answered half of our original question: (3.3) holds with $\\alpha = 2$. To show that $\\alpha = 2$ is minimal, we must show that the three properties in Theorem 3.1 are mutually incompatible when $\\alpha < 2$. Then Theorem 3.1 implies (3.3) is false for this range of $\\alpha$.\n\nOur strategy will be to prove that the Bellman function cannot possess the three properties assuming it is twice continuously differentiable. Then we will drop this assumption using a mollification argument. This scheme will first require a couple of lemmas about a class of differential inequalities. \\footnote{The author thanks Iosif Pinelis for his assistance provided via \\texttt{mathoverflow.net} in the proof of these lemmas.}\n\n\\begin{lemma}\nIf $g:\\mathbb{R} \\to \\mathbb{R}$ satisfies $g'' \\leq -bg$ where $b$ is some positive constant and $g(x_0) > 0$, then there is some $x > x_0$ such that $g(x) = 0$.\n\\end{lemma}\n\n\\begin{proof}\nOf course, by the intermediate value theorem, it suffices to show that eventually $g(x) \\leq 0$. Suppose for contradiction that $g(x) > 0$ for all $x>x_0$. Then on $[x_0,\\infty)$, $g'$ is monotonically decreasing because $g'' \\leq -bg < 0$. Thus $\\lim\\limits_{x \\to \\infty}g'(x)$ exists and is either finite or $-\\infty$. Call this limit $L$.\\\\\n\nIf $L<0$, then we are done because $g'(x)$ is eventually bounded above by $L+\\epsilon < 0$ and thus $g(x) \\to -\\infty$.\\\\\n\nOn the other hand, suppose that $L \\geq 0$. Then $g'(x)\\geq 0$ for all $x \\geq x_0$, and thus for such $x$ we have $g(x) \\geq g(x_0)$.\\\\\n\nNow fix $\\epsilon \\in \\left(0,\\frac{bg(x_0)}{2}\\right)$. Select $c$ such that $|g'(x) - L| < \\epsilon$ when $x>c$. Lastly, choose $x_1$ and $x_2$ such that $c<x_1<x_2$ and $x_2 - x_1 > 2$. Then for some $t \\in (x_1, x_2)$ we have\n\n\\begin{align*}\ng''(t) = \\frac{g'(x_2) - g'(x_1)}{x_2 - x_1} &\\geq \\frac{L-\\epsilon - (L + \\epsilon)}{x_2 - x_1}\\\\\n&= \\frac{-2\\epsilon}{x_2 - x_1}\\\\\n&\\geq -\\epsilon\\\\\n&\\geq \\frac{-bg(x_0)}{2}\n\\end{align*}\n\nTherefore,\n$$\\frac{-bg(x_0)}{2} \\leq g''(t) \\leq -bg(t) \\leq -bg(x_0)$$\nwhich gives our contradiction.\n\n\\end{proof}\n\n\\begin{lemma}\nSuppose $g:\\mathbb{R} \\to \\mathbb{R}$ satisfies $g'' \\leq -bg$ where $b$ is some real positive constant. If $g(x_0) > 0$ then $g(x) = 0$ for some $x \\in \\big[x_0, x_0 + \\frac{\\pi}{\\sqrt{b}}\\big]$.\n\\end{lemma}\n\n\\begin{proof}\nWe break into two cases.\\\\\n\nCase 1: $g'(x_0) \\leq 0$\\\\\nLet $x_1 = \\min\\{x>x_0: g(x)=0\\}$. We know $x_1$ exists by the previous lemma. Note that $g''\\leq -bg \\leq 0$ on $[x_0,x_1]$ since $g\\geq 0$ there. Hence $g'\\leq 0 $ on $ [x_0,x_1]$ because $g'(x_0)\\leq 0$ and $g'$ is decreasing.\nLet $E(x) = g'(x)^2 + bg(x)^2$. Then for all $x \\in [x_0,x_1]$\n\\begin{align*}\nE'(x) &= 2g'(x)g''(x) + 2bg(x)g'(x)\\\\\n&= 2g'(x)(g''(x) + bg(x))\\\\\n&\\geq 0\n\\end{align*}\nTherefore $E(x_0)\\leq E(x)$ for all $x\\in[x_0,x_1]$, i.e., $g'(x_0)^2 + bg(x_0)^2 \\leq g'(x)^2 + bg(x)^2$. Therefore, recalling that $g'(x)\\leq$0, we have\n$$1 \\leq \\frac{-g'(x)}{\\sqrt{g'(x_0)^2 + bg(x_0)^2 - bg(x)^2}}$$\nIntegrating both sides over the interval $[x_0,x_1]$ yields\n\\begin{align*}\nx_1 - x_0 &\\leq \\left.\\frac{-1}{\\sqrt{b}}\\sin^{-1}\\left(g(x)\\sqrt{\\frac{b}{g'(x_0)^2 + bg(x_0)^2}}\\right)\\right\\vert_{x=x_0}^{x=x_1}\\\\\n&= \\frac{-1}{\\sqrt{b}}\\left[\\sin^{-1}(0) - \\sin^{-1}\\left(\\sqrt{\\frac{bg(x_0)^2}{g'(x_0)^2 + bg(x_0)^2}}\\right)\\right]\\\\\n&\\leq \\frac{-1}{\\sqrt{b}}(0 - \\sin^{-1}(1))\\\\\n& = \\frac{\\pi}{2\\sqrt{b}}\n\\end{align*}\n\nCase 2: $g'(x_0) > 0$\\\\\nThis time we let $x_2 = \\min\\{x>x_0: g(x)=0\\}$. Again we note that $g''\\leq 0$ on $[x_0,x_2]$. Therefore, there is a point $x_1 \\in [x_0,x_2]$ such that $g'(x_1) = 0$ with $g'>0$ on $[x_0,x_1]$ and $g'<0$ on $[x_1,x_2]$. From Case 1, we know that $x_2 - x_1 \\leq \\frac{\\pi}{2\\sqrt{b}}$. It suffices to show $x_1 - x_0 \\leq \\frac{\\pi}{2\\sqrt{b}}$.\\\\\n\nRecall $E(x) = g'(x)^2 + bg(x)^2$ and $E'(x) = 2g'(x)(g''(x) + bg(x))$. Since $g' \\geq 0$ on $[x_0,x_1]$, $E'(x) \\leq 0$ there. Thus on that interval, we have $g'(x)^2 + bg(x)^2 \\geq g'(x_1)^2 + bg(x_1)^2 = bg(x_1)^2$ which implies\n$$1 \\leq \\frac{g'(x)}{\\sqrt{bg(x_1)^2 - bg(x)^2}}$$\nAs before, we integrate both sides over $[x_0,x_1]$ to get\n\\begin{align*}\nx_1 - x_0 &\\leq \\left.\\frac{1}{\\sqrt{b}}\\sin^{-1}\\left(\\frac{g(x)}{g(x_1)}\\right)\\right\\vert_{x=x_0}^{x=x_1}\\\\\n&= \\frac{1}{\\sqrt{b}}\\left[\\sin^{-1}(1) - \\sin^{-1}\\left(\\frac{g(x_0)}{g(x_1)}\\right)\\right]\\\\\n&\\leq \\frac{\\pi}{2\\sqrt{b}}\n\\end{align*}\n\n\\end{proof}\n\n\\begin{theorem}\nIf $\\mathcal{B}(x,0) \\in \\mathcal{C}^2(\\mathbb{R})$, then $\\alpha = 2$ is the minimal constant such that (3.3) holds. Here $\\mathcal{B}$ is the theoretical Bellman function as in the the proof of Theorem 3.1.\n\\end{theorem}\n\n\\begin{proof}\nRecall that $\\mathcal{B}(x,y) = \\sup\\{\\mathbb{E}V(f_n,(Sf)^2_n) : (f_n,(Sf)^2_n)\\in \\mathcal{M}(x,y)\\}$ where $\\mathcal{M}(x,y)$ is the set of all $\\mathbb{R} \\times [0,\\infty)$ valued processes $(f_n,(Sf)^2_n)$ such that $\\{f_n\\}$ is a dyadic martingale with $f_0 = x$ and $(Sf)^2_n = y + \\sum_{k=1}^n (df_k)^2$. By the proof of Theorem 3.1, $\\mathcal{B}$ satisfies the majorization and concavity properies as well as the initial condition as long as (3.3) holds. We will show that if $\\alpha < 2$, $\\mathcal{B}$ does not obey these properties, and thus (3.3) is false.\n\nFrom the definition of $\\mathcal{B}$, we have $$e^{\\alpha y}\\mathcal{B}(x - \\alpha y, 0) = \\mathcal{B}(x,y)$$\nThis is essentially a consequence of the fact that $V$ has this property. Thus, as before, if we define $f(x) = \\mathcal{B}(x,0)$ then $\\mathcal{B}$ has the form\n\\begin{equation}\n\\mathcal{B}(x,y) = e^{\\alpha y}f(x - \\alpha y)\n\\end{equation}\nPlugging (3.13) into the concavity property and setting $y=0$ gives\n\\begin{equation}\n\\frac{f(x - \\delta -\\alpha \\delta^2) + f(x + \\delta -\\alpha \\delta^2)}{2} \\leq e^{-\\alpha\\delta^2}f(x)\n\\end{equation}\nIf we express $f$ as a second order Taylor polynomial in $\\delta$ centered at $x$ (valid because we are assuming $f\\in\\mathcal{C}^2$), then (3.14) becomes\n\\begin{align*}\nf(x) - \\alpha\\delta^2 f'(x) + \\frac{\\delta^2}{2} f''(x) + O(\\delta^3) &\\leq e^{-\\alpha\\delta^2}f(x)\\\\\n&= f(x) - \\alpha\\delta^2 f(x) + O(\\delta^4)\n\\end{align*}\nDividing through by $\\frac{\\delta^2}{2}$ and letting $\\delta \\to 0$, we get\n\\begin{equation}\nf''(x) - 2\\alpha f'(x) +2\\alpha f(x) \\leq 0\n\\end{equation}\nMultiply both sides by $e^{-\\alpha x}$ to get\n\\begin{equation*}\ne^{-\\alpha x}f''(x) - 2\\alpha e^{-\\alpha x}f'(x) +2\\alpha e^{-\\alpha x}f(x) = \\left(e^{-\\alpha x}f(x) \\right)'' + (2\\alpha - \\alpha^2)e^{-\\alpha x}f(x) \\leq 0\n\\end{equation*}\nLetting $g(x) \\coloneqq e^{-\\alpha x}f(x)$ and $b \\coloneqq 2\\alpha - \\alpha^2$ we have $g'' + bg \\leq 0$.\n\nSuppose\\footnote{Assuming $\\alpha > 0$ is valid because if (3.3) fails for some $\\alpha$, it clearly fails for all smaller $\\alpha$.} $0 < \\alpha < 2$ so that $b > 0$. Then by Lemma 3.3, $g(x) \\leq 0$  and thus $f(x) \\leq 0$ for some $x>0$. Hence, for this $x$, $\\mathcal{B}(x,0) = f(x) \\leq 0 < V(x,0) = e^x - 1$. Therefore, for this range of $\\alpha$, the Bellman function $\\mathcal{B}$ cannot posses both the majorization and concavity properties of Theorem 3.1, and (3.3) is false.\n\n\\end{proof}\n\nThe next theorem allows us to drop the smoothness assumption on $\\mathcal{B}$. However, we will use a mollification argument which requires $\\mathcal{B}$ to be continuous. We will subsequently show that the continuity of $\\mathcal{B}$ is a consequence of the concavity property.\n\n\\begin{theorem}\nIf $\\mathcal{B}(x,0)$ is continuous, then $\\alpha = 2$ is the minimal constant such that (3.3) holds.\n\\end{theorem}\n\n\\begin{proof}\nLet $\\eta(x)$ be the standard mollifier.\n$$\n\\eta(x) \\coloneqq \\begin{cases}\nCe^{\\frac{1}{x^2-1}} & \\text{if } |x| < 1\\\\\n0 & \\text{if } |x| \\geq 1\n\\end{cases}\n$$\nwith $C>0$ chosen so that $\\int_\\mathbb{R} \\eta(x) dx = 1$.\nLet $\\eta_\\epsilon(x) \\coloneqq \\frac{1}{\\epsilon}\\eta(\\frac{x}{\\epsilon})$.\nConvolving both sides of (3.14) with $\\eta_\\epsilon$ gives\n$$\\frac{f_\\epsilon(x - \\delta -\\alpha \\delta^2) + f_\\epsilon(x + \\delta -\\alpha \\delta^2)}{2} \\leq e^{-\\alpha\\delta^2}f_\\epsilon(x)$$\nwhere $f_\\epsilon = f*\\eta_\\epsilon$.\n\n$f_\\epsilon \\in C^\\infty(\\mathbb{R})$, so by the proof of the previous theorem, we have $g_\\epsilon'' + bg_\\epsilon \\leq 0$ where $g_\\epsilon(x) \\coloneqq e^{-\\alpha x}f_\\epsilon(x)$ and $b \\coloneqq 2\\alpha - \\alpha^2$. As before, we suppose $0 < \\alpha < 2$ so that $b>0$.\n\nFix $\\epsilon' > 0$. Select $x_0$ such that $\\epsilon' < e^{x_0} - 1$. Since $f$ is continuous by assumption, $f_\\epsilon \\to f$ uniformly on compact sets \\cite{Evans}, and we can select $\\epsilon$ such that $\\left \\vert f - f_\\epsilon \\right \\vert < \\epsilon'$ on $[x_0,x_0 + \\frac{\\pi}{\\sqrt{b}}]$. By Lemma 3.4, there exists $x \\in [x_0, x_0 + \\frac{\\pi}{\\sqrt{b}}]$ such that $g_\\epsilon(x) \\leq 0$ which implies $f_\\epsilon(x) \\leq 0$. Therefore, for this $x$, we have\n$$\\mathcal{B}(x,0) = f(x) < f_\\epsilon(x) + \\epsilon' \\leq \\epsilon' < e^{x_0}-1<e^x-1 = V(x,0)$$\n\nOnce again, we've shown $\\mathcal{B}$ cannot posses both the majorization and concavity properties of Theorem 3.1, and so (3.3) is false for $\\alpha < 2$ assuming $\\mathcal{B}$ is continuous.\n\\end{proof}\n\nOur final task is to justify the assumption that $\\mathcal{B}$ is continuous. Indeed, this follows from the concavity property.\n\n\\begin{lemma}\nGiven $\\alpha \\in \\mathbb{R}$, if $f:\\mathbb{R} \\to \\mathbb{R}$ is an increasing function such that\n$$\\frac{f(x_0 - t -\\alpha t^2) + f(x_0 + t -\\alpha t^2)}{2} \\leq e^{-\\alpha t^2}f(x_0)$$\nfor all $t$, then $f$ is continuous at $x_0$.\n\\end{lemma}\n\n\\begin{proof}\nLet $U = \\lim\\limits_{x \\to x_0^+} f(x)$ and $L = \\lim\\limits_{x \\to x_0^-} f(x)$. $U$ and $L$ exists with $L \\leq U$ because $f$ is increasing. It suffices to show that $L \\geq U$.\n\nFix $\\epsilon > 0$. Choose $\\delta > 0$ such that $f(x) > L - \\epsilon$ when $x \\in (x_0 - \\delta, x_0)$. Select $t$ such that\n$$0 < t < \\min\\left(\\frac{1}{\\alpha},\\sqrt{\\frac{\\delta}{\\alpha} + \\frac{1}{4}}-\\frac{1}{2}\\right)$$\nNote that this selection of $t$ ensures that $0 < t - \\alpha t^2$ and $\\alpha t^2 + t < \\delta$. Finally, choose $x$ such that\n$$\\min (x_0 + \\alpha t^2 - t, x_0 + \\alpha t^2 + t - \\delta) < x < x_0$$\nSuch an $x$ is known to exists since $\\alpha t^2 - t < 0$ and $\\alpha t^2 + t < \\delta$.\nThen we have\n$$x_0 - \\delta < x - \\alpha t^2 - t < x < x_0 < x - \\alpha t^2 + t$$\nand hence\n$$\\frac{L - \\epsilon + U}{2} \\leq \\frac{f(x - \\alpha t^2 - t) + f(x - \\alpha t^2 + t)}{2} \\leq f(x)e^{-\\alpha t} \\leq f(x) \\leq L$$\nSince $\\epsilon$ was arbitary, we have $\\frac{L + U}{2} \\leq L$ and thus $U \\leq L$ as desired.\n\\end{proof}\n\nWe are now prepared to state our main result.\n\n\\begin{theorem}\n$\\alpha = 2$ is the smallest constant such that\n$$\\mathbb{E}e^{f_n} \\leq \\mathbb{E}e^{\\alpha(Sf_n)^2}$$\nfor all dyadic martingales with $f_0 = 0$.\n\\end{theorem}\n\n\\begin{proof}\nIt is clear from the definition that $\\mathcal{B}(x,0)$ is increasing. This fact coupled with the concavity property implies $\\mathcal{B}(x,0)$ is continuous by the previous lemma. Now apply Theorem 3.6.\n\\end{proof}\n\n\\section{Examples and Future Work}\nIt is worth noting that a simpler proof of Theorem 3.2 exists. It uses a result known as Rubin's lemma \\cite{Pipher}, which says that if $\\{f_n\\}$ is a real valued dyadic martingale on $[0,1)$ whose limit is $f$, then for all $\\lambda \\geq 0$,\n$$\\int_0^1 e^{\\lambda\\left( f(x) -\\langle f \\rangle_{[0,1)}\\right) - \\frac{\\lambda^2}{2}\\left(Sf\\right)^2(x)}dx \\leq 1$$\nSee \\cite{BM} for a proof.\n\nWe can use the Cauchy-Schwarz inequality with Rublin's lemma to prove Theorem 3.2. As usual, we may work on $([0,1),\\mathcal{B},|\\cdot|)$.\n\\begin{align*}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx &= \\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}-(Sf)^2(x)+(Sf)^2(x)}dx &\\\\\n&\\leq \\sqrt{\\int_0^1 e^{2[f(x)-\\langle f \\rangle_{[0,1)}]-2(Sf)^2(x)}dx} \\sqrt{\\int_0^1 e^{2(Sf)^2(x)}dx} &\\\\\n&\\leq \\sqrt{1} \\sqrt{\\int_0^1 e^{2(Sf)^2(x)}dx} &\\\\\n&\\leq \\int_0^1 e^{2 (Sf)^2(x)}dx &\n\\end{align*}\n\nThe Bellman function method is still desirable in at least two ways. First, it's a general technique for proving inequalities, while this simpler proof is very particular to the details of our problem. Second, the Bellman function allowed us to prove the sharpness of Theorem 3.2. The shorter proof would requiring an accompanying construction of a martingale which maximizes the left side of the inequality relative to the right side.\n\nSuch a construction could provide the basis for some future work. Sometimes, extremal examples can be deduced from the Bellman function itself (see, e.g. \\cite{Wang}). Although we were unable to do this, we have an example of a martingale which shows that\n\\begin{equation}\n\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\leq \\int_0^1 e^{\\alpha (Sf)^2(x)}dx\n\\end{equation}\nis false for $\\alpha < \\log_2(e) \\approx 1.44$. Recall that $\\alpha = 2$ was optimal.\n\nLet $f=\\sum\\limits_{n=0}^\\infty \\chi_{I_n^0} = \\sum\\limits_{n=1}^\\infty n\\chi_{I_n^1}$ where $I_n^k = \\left[\\frac{k}{2^n},\\frac{k+1}{2^n}\\right)$. As always, we can think of $f$ as a dyadic martingale by letting $f_n = \\mathbb{E}[f\\vert \\mathcal{D}_n]$ where $\\mathcal{D}_n$ is the $\\sigma$-algebra generated by the nth generation of dyadic subintervals of $[0,1)$. This function is a discrete approximation of $-\\log_2x$ has the property that $(Sf)^2 = f$. It is a \"fixed point\" of the operator that sends $f \\mapsto (Sf)^2$. The left side of (4.1) becomes\n$$\\int_0^1 e^{f(x)-\\langle f \\rangle_{[0,1)}}dx \\approx \\int_0^1 e^{-\\log_2x}dx = \\int_0^1 x^{-\\log_2(e)}dx = \\infty$$\nOn the other hand, the right side becomes\n$$\\int_0^1 e^{\\alpha (Sf)^2(x)}dx = \\int_0^1 e^{\\alpha f(x)}dx \\approx \\int_0^1 e^{-\\alpha\\log_2x}dx = \\int_0^1 x^{-\\alpha\\log_2(e)}dx$$\nwhich is finite for $\\alpha < \\frac{1}{\\log_2(e)} = \\ln(2) \\approx .69$. So our example falsifies (4.1) for this range of $\\alpha$.\n\nWe can push this example further. $[S(\\lambda g)]^2 = \\lambda^2 [S(g)]^2$ for all $g$ follows immediately from the definition of $S$. Applying this to our present example gives $[S(\\lambda f)]^2 = \\lambda^2 [S(f)]^2 = \\lambda^2 f$. Plugging $\\lambda f$ into (4.1), the left side is approximately $\\int_0^1 x^{-\\lambda\\log_2(e)}dx$ and the right side is approximately $\\int_0^1 x^{-\\alpha\\lambda^2\\log_2(e)}dx$. Thus for $\\lambda = \\frac{1}{\\log_2(e)}$, the left side is infinite and the right side is finite when $\\alpha < \\log_2(e) \\approx 1.44$, so $\\frac{f}{\\log_2(e)}$ falsifies (4.1) for this range of $\\alpha$.\n\nWe are still left without a martingale showing $\\alpha = 2$ is sharp. While the Bellman function proof makes it unnecessary, finding an explicit example is an interesting future problem.\n\\bibliographystyle{ieeetr}\n", "meta": {"timestamp": "2018-05-29T02:06:56", "yymm": "1805", "arxiv_id": "1805.10510", "language": "en", "url": "https://arxiv.org/abs/1805.10510", "abstract": "We describe the Bellman function technique for proving sharp inequalities in harmonic analysis. To provide an example along with historical context, we present how it was originally used by Donald Burkholder to prove $L^p$ boundedness of the $\\pm 1$ martingale transform. Finally, with Burkholder's result as a blueprint, we use the Bellman function to prove a new result related to the Chang-Wilson-Wolff Inequality.", "subjects": "Classical Analysis and ODEs (math.CA)", "title": "A modification of the Chang-Wilson-Wolff Inequality via the Bellman Function", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9848109516378094, "lm_q2_score": 0.8558511396138365, "lm_q1q2_score": 0.842851575263406}}
{"url": "https://arxiv.org/abs/2009.07893", "title": "Largest small polygons: A sequential convex optimization approach", "abstract": "A small polygon is a polygon of unit diameter. The maximal area of a small polygon with $n=2m$ vertices is not known when $m\\ge 7$. Finding the largest small $n$-gon for a given number $n\\ge 3$ can be formulated as a nonconvex quadratically constrained quadratic optimization problem. We propose to solve this problem with a sequential convex optimization approach, which is an ascent algorithm guaranteeing convergence to a locally optimal solution. Numerical experiments on polygons with up to $n=128$ sides suggest that the optimal solutions obtained are near-global. Indeed, for even $6 \\le n \\le 12$, the algorithm proposed in this work converges to known global optimal solutions found in the literature.", "text": "\\section{Introduction}\nThe {\\em diameter} of a polygon is the largest Euclidean distance between pairs of its vertices. A polygon is said to be {\\em small} if its diameter equals one. For a given integer $n \\ge 3$, the maximal area problem consists in finding the small $n$-gon with the largest area. The problem was first investigated by Reinhardt~\\cite{reinhardt1922} in 1922. He proved that\n\\begin{itemize}\n\\item when $n$ is odd, the regular small $n$-gon is the unique optimal solution;\n\\item when $n=4$, there are infinitely many optimal solutions, including the small square;\n\\item when $n \\ge 6$ is even, the regular small $n$-gon is not optimal.\n\\end{itemize}\n\nThe maximal area is known for even $n \\le 12$. In 1961, Bieri~\\cite{bieri1961} found the largest small $6$-gon, assuming the existence of an axis of symmetry. In 1975, Graham~\\cite{graham1975} independently constructed the same $6$-gon, represented in Figure~\\ref{figure:6gon:U6}. In 2002, Audet, Hansen, Messine, and Xiong~\\cite{audet2002} combined Graham's strategy with global optimization methods to find the largest small $8$-gon, illustrated in Figure~\\ref{figure:8gon:U8}. In 2013, Henrion and Messine~\\cite{henrion2013} found the largest small $10$- and $12$-gons by also solving globally a nonconvex quadratically constrained quadratic optimization problem. They also found the largest small axially symmetrical $14$- and $16$-gons. In 2017, Audet~\\cite{audet2017} showed that the regular small polygon has the maximal area among all equilateral small polygons. In 2020, Audet, Hansen, and Svrtan~\\cite{audet2020} determined analytically the largest small axially symmetrical $8$-gon.\n\nThe diameter graph of a small polygon is defined as the graph with the vertices of the polygon, and an edge between two vertices exists only if the distance between these vertices equals one. Graham~\\cite{graham1975} conjectured that, for even $n \\ge 6$, the diameter graph of a small $n$-gon with maximal area has a cycle of length $n-1$ and one additional edge from the remaining vertex. The case $n=6$ was proven by Graham himself~\\cite{graham1975} and the case $n=8$ by Audet, Hansen, Messine, and Xiong~\\cite{audet2002}. In 2007, Foster and Szabo~\\cite{foster2007} proved Graham's conjecture for all even $n \\ge 6$. Figure~\\ref{figure:4gon}, Figure~\\ref{figure:6gon}, and Figure~\\ref{figure:8gon} show diameter graphs of some small polygons. The solid lines illustrate pairs of vertices which are unit distance apart.\n\nIn addition to exact results and bounds, uncertified largest small polygons have been obtained both by metaheurisitics and nonlinear optimization. Assuming Graham's conjecture and the existence of an axis of symmetry, Mossinghoff~\\cite{mossinghoff2006b} in 2006 constructed large small $n$-gons for even $6 \\le n \\le 20$. In 2018, using a formulation based on polar coordinates, Pinter~\\cite{pinter2018} presented numerical solutions estimates of the maximal area for even $6 \\le n \\le 80$. However, the solutions obtained by Pinter are not optimal for $n\\ge 32$.\n\nThe maximal area problem can be formulated as a nonconvex quadratically constrained quadratic optimization problem. In this work, we propose to solve it with a sequential convex optimization approach, also known as the concave-convex procedure~\\cite{marks1978,lanckriet2009}. This approach is an ascent algorithm guaranteeing convergence to a locally optimal solution. Numerical experiments on polygons up to $n=128$ sides suggest that the optimal solutions obtained are near-global. Indeed, without assuming Graham's conjecture nor the existence of an axis of symmetry in our quadratic formulation, optimal $n$-gons obtained with the algorithm proposed in this work verify both conditions within the limit of numerical computations. Moreover, for even $6 \\le n \\le 12$, this algorithm converges to known global optimal solutions. The algorithm is implemented as a MATLAB-based package, OPTIGON, which is available on GitHub~\\cite{optigon}. OPTIGON requires that CVX~\\cite{cvx2} be installed.\n\nThe remainder of this paper is organized as follows. In Section~\\ref{sec:ngon}, we recall principal results on largest small polygons. Section~\\ref{sec:nqcqo} presents the quadratic formulation of the maximal area problem and the sequential convex optimization approach to solve it. We report in Section~\\ref{sec:results} computational results. Section~\\ref{sec:conclusion} concludes the paper.\n\n\\begin{figure}[h]\n\t\\centering\n\t\\subfloat[$(\\geo{R}_4,0.5)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.5000,0.5000) -- (0,1) -- (-0.5000,0.5000) -- cycle;\n\t\t\\draw (0,0) -- (0,1);\n\t\t\\draw (0.5000,0.5000) -- (-0.5000,0.5000);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{R}_3^+,0.5)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0.5000,0.8660) -- (0,1) -- (-0.5000,0.8660);\n\t\t\\draw (0,1) -- (0,0) -- (0.5000,0.8660) -- (-0.5000,0.8660) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\caption{Two small $4$-gons $(\\geo{P}_4,A(\\geo{P}_4))$}\n\t\\label{figure:4gon}\n\\end{figure}\n\n\\begin{figure}[h]\n\t\\centering\n\t\\subfloat[$(\\geo{R}_6,0.649519)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.4330,0.2500) -- (0.4330,0.7500) -- (0,1) -- (-0.4330,0.7500) -- (-0.4330,0.2500) -- cycle;\n\t\t\\draw (0,0) -- (0,1);\n\t\t\\draw (0.4330,0.2500) -- (-0.4330,0.7500);\n\t\t\\draw (0.4330,0.7500) -- (-0.4330,0.2500);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{R}_5^+,0.672288)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.5000,0.3633) -- (0.3090,0.9511) -- (0,1) -- (-0.3090,0.9511) -- (-0.5000,0.3633) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.3090,0.9511) -- (-0.5000,0.3633) -- (0.5000,0.3633) -- (-0.3090,0.9511) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{U}_6,0.674981)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.5000,0.4024) -- (0.3438,0.9391) -- (0,1) -- (-0.3438,0.9391) -- (-0.5000,0.4024) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.3438,0.9391) -- (-0.5000,0.4024) -- (0.5000,0.4024) -- (-0.3438,0.9391) -- (0,0);\n\t\t\\end{tikzpicture}\n\t\t\\label{figure:6gon:U6}\n\t}\n\t\\caption{Three small $6$-gons $(\\geo{P}_6,A(\\geo{P}_6))$}\n\t\\label{figure:6gon}\n\\end{figure}\n\n\\begin{figure}[h]\n\t\\centering\n\t\\subfloat[$(\\geo{R}_8,0.707107)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.3536,0.1464) -- (0.5000,0.5000) -- (0.3536,0.8536) -- (0,1) -- (-0.3536,0.8536) -- (-0.5000,0.5000) -- (-0.3536,0.1464) -- cycle;\n\t\t\\draw (0,0) -- (0,1);\n\t\t\\draw (0.3536,0.1464) -- (-0.3536,0.8536);\n\t\t\\draw (0.5000,0.5000) -- (-0.5000,0.5000);\n\t\t\\draw (0.3536,0.8536) -- (-0.3536,0.1464);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{R}_7^+,0.725320)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.4010,0.1931) -- (0.5000,0.6270) -- (0.2225,0.9749) -- (0,1) -- (-0.2225,0.9749) -- (-0.5000,0.6270) -- (-0.4010,0.1931) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.2225,0.9749) -- (-0.4010,0.1931) -- (0.5000,0.6270) -- (-0.5000,0.6270) -- (0.4010,0.1931) -- (-0.2225,0.9749) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{U}_8,0.726868)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.4091,0.2238) -- (0.5000,0.6404) -- (0.2621,0.9650) -- (0,1) -- (-0.2621,0.9650) -- (-0.5000,0.6404) -- (-0.4091,0.2238) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.2621,0.9650) -- (-0.4091,0.2238) -- (0.5000,0.6404) -- (-0.5000,0.6404) -- (0.4091,0.2238) -- (-0.2621,0.9650) -- (0,0);\n\t\t\\end{tikzpicture}\n\t\t\\label{figure:8gon:U8}\n\t}\n\t\\caption{Three small $8$-gons $(\\geo{P}_8,A(\\geo{P}_8))$}\n\t\\label{figure:8gon}\n\\end{figure}\n\n\\section{Largest small polygons}\\label{sec:ngon}\nLet $A(\\geo{P})$ denote the area of a polygon $\\geo{P}$. Let $\\geo{R}_n$ denote the regular small $n$-gon. We have\n\\[\nA(\\geo{R}_n) =\n\\begin{cases}\n\\frac{n}{2}\\left(\\sin \\frac{\\pi}{n} - \\tan \\frac{\\pi}{2n}\\right) &\\text{if $n$ is odd,}\\\\\n\\frac{n}{8}\\sin \\frac{2\\pi}{n} &\\text{if $n$ is even.}\\\\\n\\end{cases}\n\\]\nWe remark that $A(\\geo{R}_n) < A(\\geo{R}_{n-1})$ for all even $n\\ge 6$~\\cite{audet2009}. This suggests that $\\geo{R}_n$ does not have maximum area for any even $n\\ge 6$. Indeed, when $n$ is even, we can construct a small $n$-gon with a larger area than $\\geo{R}_n$ by adding a vertex at distance $1$ along the mediatrix of an angle in $\\geo{R}_{n-1}$. We denote this $n$-gon by $\\geo{R}_{n-1}^+$ and we have\n\\[\nA(\\geo{R}_{n-1}^+) = \\frac{n-1}{2} \\left(\\sin \\frac{\\pi}{n-1} - \\tan \\frac{\\pi}{2n-2}\\right) + \\sin \\frac{\\pi}{2n-2} - \\frac{1}{2}\\sin \\frac{\\pi}{n-1}.\n\\]\n\n\\begin{theorem}[Reinhardt~\\cite{reinhardt1922}]\nFor all $n \\ge 3$, let $A_n^*$ denote the maximal area among all small $n$-gons and let $\\ub{A}_n := \\frac{n}{2}\\left(\\sin \\frac{\\pi}{n} - \\tan \\frac{\\pi}{2n}\\right)$.\n\\begin{itemize}\n\\item When $n$ is odd, $A_n^* = \\ub{A}_n$ is only achieved by $\\geo{R}_n$.\n\\item $A_4^* = 0.5 < \\ub{A}_4$ is achieved by infinitely many $4$-gons, including $\\geo{R}_4$ and~$\\geo{R}_3^+$ illustrated in Figure~\\ref{figure:4gon}.\n\\item When $n\\ge 6$ is even, $A(\\geo{R}_n) < A_n^* < \\ub{A}_n$.\n\\end{itemize}\n\\end{theorem}\n\nThe maximal area~$A_n^*$ is known for even $n \\le 12$. Using geometric arguments, Graham~\\cite{graham1975} determined analytically the largest small $6$-gon, represented in Figure~\\ref{figure:6gon:U6}. Its area $A_6^* \\approx 0.674981$ is about $3.92\\%$ larger than $A(\\geo{R}_6) \\approx 0.649519$. The approach of Graham, combined with methods of global optimization, has been followed by~\\cite{audet2002} to determine the largest small $8$-gon, represented in Figure~\\ref{figure:8gon:U8}. Its area $A_8^* \\approx 0.726868$ is about $2.79\\%$ larger than $A(\\geo{R}_8) \\approx 0.707107$. Henrion and Messine~\\cite{henrion2013} found that $A_{10}^* \\approx 0.749137$ and $A_{12}^* \\approx 0.760730$.\n\nFor all even $n\\ge 6$, let $\\geo{U}_n$ denote the largest small $n$-gon.\n\\begin{theorem}[Graham~\\cite{graham1975}, Foster and Szabo~\\cite{foster2007}]\n\\label{thm:area:diam}\nFor even $n \\ge 6$, the diameter graph of $\\geo{U}_n$ has a cycle of length $n-1$ and one additional edge from the remaining vertex.\n\\end{theorem}\n\\begin{conjecture}\n\\label{thm:area:sym}\nFor even $n \\ge 6$, $\\geo{U}_n$ has an axis of symmetry corresponding to the pending edge in its diameter graph.\n\\end{conjecture}\nFrom Theorem~\\ref{thm:area:diam}, we note that $\\geo{R}_{n-1}^+$ has the same diameter graph as the largest small $n$-gon $\\geo{U}_n$. Conjecture~\\ref{thm:area:sym} is only proven for $n=6$ and this is due to Yuan~\\cite{yuan2004}. However, the largest small polygons obtained by~\\cite{audet2002} and~\\cite{henrion2013} are a further evidence that the conjecture may be true.\n\n\\section{Nonconvex quadratically constrained quadratic optimization} \\label{sec:nqcqo}\nWe use cartesian coordinates to describe an $n$-gon $\\geo{P}_n$, assuming that a vertex $\\geo{v}_i$, $i=0,1,\\ldots,n-1$, is positioned at abscissa $x_i$ and ordinate $y_i$. Placing the vertex $\\geo{v}_0$ at the origin, we set $x_0 = y_0 = 0$. We also assume that the $n$-gon $\\geo{P}_n$ is in the half-plane $y\\ge 0$ and the vertices $\\geo{v}_i$, $i=1,2,\\ldots,n-1$, are arranged in a counterclockwise order as illustrated in Figure~\\ref{figure:model}, i.e., $y_{i+1}x_i \\ge x_{i+1}y_i$ for all $i=1,2,\\ldots,n-2$. The maximal area problem can be formulated as follows\n\\begin{subequations}\\label{eq:ngon:area}\n\t\\begin{align}\n\t\\max_{\\rv{x},\\rv{y},\\rv{u}} \\quad & \\sum_{i=1}^{n-2} u_i\\\\\n\t\\subj \\quad & (x_j - x_i)^2 + (y_j - y_i)^2 \\le 1 &\\forall 1\\le i < j \\le n-1,\\label{eq:ngon:d}\\\\\n\t& x_i^2 + y_i^2 \\le 1 &\\forall 1 \\le i \\le n-1,\\label{eq:ngon:r}\\\\\n\t& y_i \\ge 0 &\\forall 1 \\le i \\le n-1,\\label{eq:ngon:y}\\\\\n\t& 2u_i \\le y_{i+1}x_i - x_{i+1}y_i &\\forall 1 \\le i \\le n-2,\\label{eq:ngon:u}\\\\\n\t& u_i \\ge 0 &\\forall 1 \\le i \\le n-2.\n\t\\end{align}\n\\end{subequations}\nAt optimality, for all $i=1,2,\\ldots,n-2$, $u_i = (y_{i+1}x_i - x_{i+1}y_i)/2$, which corresponds to the area of the triangle $\\geo{v}_0\\geo{v}_i\\geo{v}_{i+1}$.\nIt is important to note that, unlike what was done in~\\cite{audet2002,henrion2013}, this formulation does not make the assumption of Graham's conjecture, nor of the existence of an axis of symmetry.\n\n\\begin{figure}[h]\n\\centering\n\\begin{tikzpicture}[scale=4]\n\\draw[dashed] (0,0) node[below]{$\\geo{v}_0(0,0)$} -- (0.4370,0.4370) node[right]{$\\geo{v}_1(x_1,y_1)$} -- (0.1564,0.9877) node[right]{$\\geo{v}_2(x_2,y_2)$} -- (-0.1564,0.9877) node[above]{$\\geo{v}_3(x_3,y_3)$} -- (-0.4540,0.8911) node[left]{$\\geo{v}_4(x_4,y_4)$} -- (-0.5507,0.2806) node[left]{$\\geo{v}_5(x_5,y_5)$} -- cycle;\n\\draw[->] (-0.25,0)--(0.25,0)node[below]{$x$};\n\\draw[->] (0,0)--(0,0.5)node[left]{$y$};\n\\end{tikzpicture}\n\\caption{Definition of variables: Case of $n=6$ vertices}\n\\label{figure:model}\n\\end{figure}\n\nProblem~\\eqref{eq:ngon:area} is a nonconvex quadratically constrained quadratic optimization problem and can be reformulated as a difference-of-convex optimization (DCO) problem of the form\n\\begin{subequations}\\label{eq:dco}\n\t\\begin{align}\n\t\\max_{\\rv{z}} \\quad & g_0(\\rv{z}) - h_0(\\rv{z})\\\\\n\t\\subj \\quad& g_i(\\rv{z}) - h_i(\\rv{z}) \\ge 0 &\\forall 1 \\le i \\le m,\n\t\\end{align}\n\\end{subequations}\nwhere $g_0,\\ldots,g_m$ and $h_0,\\ldots,h_m$ are convex quadratic functions. We note that the feasible set\n\\[\n\\Omega := \\{\\rv{z} \\colon g_i(\\rv{z}) - h_i(\\rv{z}) \\ge 0, i =1,2,\\ldots,m\\}\n\\]\nis compact with a nonempty interior, which implies that $g_0(\\rv{z}) - h_0(\\rv{z}) < \\infty$ for all $\\rv{z} \\in \\Omega$.\n\nFor a fixed $\\rv{c}$, we have $\\lb{g}_i(\\rv{z};\\rv{c}) := g_i(\\rv{c}) + \\nabla g_i(\\rv{c})^T (\\rv{z} - \\rv{c}) \\le g_i(\\rv{z})$ for all $i=0,1,\\ldots,m$. Then the following problem\n\\begin{subequations}\\label{eq:dcocvx}\n\t\\begin{align}\n\t\\max_{\\rv{z}} \\quad & \\lb{g}_0(\\rv{z};\\rv{c}) - h_0(\\rv{z})\\\\\n\t\\subj \\quad& \\lb{g}_i(\\rv{z};\\rv{c}) - h_i(\\rv{z}) \\ge 0 &\\forall 1 \\le i \\le m\n\t\\end{align}\n\\end{subequations}\nis a convex restriction of the DCO problem~\\eqref{eq:dco} as stated by Proposition~\\ref{thm:cvxrestr}. Constraint~\\eqref{eq:ngon:u} is equivalent to\n\\[\n(y_{i+1}-x_i)^2+(x_{i+1}+y_i)^2+8u_i \\le (y_{i+1}+x_i)^2+(x_{i+1}-y_i)^2\n\\]\nfor all $i=1,2,\\ldots,n-2$. For a fixed $(\\rv{a},\\rv{b}) \\in \\mathbb{R}^{n-1} \\times \\mathbb{R}^{n-1}$, if we replace~\\eqref{eq:ngon:u} in~\\eqref{eq:ngon:area} by\n\\[\n(y_{i+1}-x_i)^2+(x_{i+1}+y_i)^2+8u_i \\le 2(b_{i+1}+a_i)(y_{i+1}+x_i)-(b_{i+1}+a_i)^2\n+ 2(a_{i+1}-b_i)(x_{i+1}-y_i)-(a_{i+1}-b_i)^2\n\\]\nfor all $i=1,2,\\ldots,n-2$, we obtain a convex restriction of the maximal area problem.\n\n\\begin{proposition}\\label{thm:cvxrestr}\nIf $\\rv{z}$ is a feasible solution of~\\eqref{eq:dcocvx} then $\\rv{z}$ is a feasible solution of~\\eqref{eq:dco}.\n\\end{proposition}\n\\begin{proof}\nLet $\\rv{z}$ be a feasible solution of~\\eqref{eq:dcocvx}, i.e., $\\lb{g}_i(\\rv{z};\\rv{c}) - h_i(\\rv{z}) \\ge 0$ for all $i=1,2,\\ldots,m$. Then $g_i(\\rv{z}) - h_i(\\rv{z}) \\ge \\lb{g}_i(\\rv{z};\\rv{c}) - h_i(\\rv{z}) \\ge 0$ for all $i=1,2,\\ldots,m$. Thus, $\\rv{z}$ is a feasible solution of~\\eqref{eq:dco}.\n\\end{proof}\n\n\\begin{proposition}\\label{thm:ascent}\nIf $\\rv{c}$ is a feasible solution of~\\eqref{eq:dco} then \\eqref{eq:dcocvx} is a feasible problem. Moreover, if $\\rv{z}^*$ is an optimal solution of~\\eqref{eq:dcocvx} then $g_0(\\rv{c}) - h_0(\\rv{c}) \\le g_0(\\rv{z}^*) - h_0(\\rv{z}^*)$.\n\\end{proposition}\n\\begin{proof}\nLet $\\rv{c}$ be a feasible solution of~\\eqref{eq:dcocvx}, i.e., $g_i(\\rv{c}) - h_i(\\rv{c}) \\ge 0$ for all $i=1,2,\\ldots,m$. Then there exists $\\rv{z} = \\rv{c}$ such that $\\lb{g}_i(\\rv{c};\\rv{c}) - h_i(\\rv{c}) = g_i(\\rv{c}) - h_i(\\rv{c}) \\ge 0$ for all $i=1,2,\\ldots,m$. Thus, \\eqref{eq:dcocvx} is a feasible problem. Moreover, if $\\rv{z}^*$ is an optimal solution of~\\eqref{eq:dcocvx}, we have $g_0(\\rv{c}) - h_0(\\rv{c}) = \\lb{g}_0(\\rv{c};\\rv{c}) - h_0(\\rv{c}) \\le \\lb{g}_0(\\rv{z}^*;\\rv{c}) - h_0(\\rv{z}^*) \\le g_0(\\rv{z}^*) - h_0(\\rv{z}^*)$.\n\\end{proof}\n\nFrom Proposition~\\ref{thm:ascent}, the optimal small $n$-gon $(\\rv{x},\\rv{y})$ obtained by solving a convex restriction of Problem~\\eqref{eq:ngon:area} constructed around a small $n$-gon $(\\rv{a},\\rv{b})$ has a larger area than this one. Proposition~\\ref{thm:local} states that if $(\\rv{a},\\rv{b})$ is the optimal $n$-gon of the convex restriction constructed around itself, then it is a local optimal $n$-gon for the maximal area problem.\n\n\\begin{proposition}\\label{thm:local}\nLet $\\rv{c}$ be a feasible solution of~\\eqref{eq:dco}. We suppose that $\\lb{\\Omega}(\\rv{c}) := \\{\\rv{z}\\colon \\lb{g}_i(\\rv{z};\\rv{c}) - h_i(\\rv{z}) \\ge 0, i =1,2,\\ldots,m\\}$ satisfies Slater condition. If $\\rv{c}$ is an optimal solution of~\\eqref{eq:dcocvx} then $\\rv{c}$ is a critical point of~\\eqref{eq:dco}.\n\\end{proposition}\n\n\\begin{proof}\nIf $\\rv{c}$ is an optimal solution of~\\eqref{eq:dcocvx} then there exist $m$ scalars $\\mu_1, \\mu_2,\\ldots,\\mu_m$ such that\n\\[\n\\begin{aligned}\n\\nabla\\lb{g}_0 (\\rv{c};\\rv{c}) + \\sum_{i=1}^m \\mu_i\\nabla \\lb{g}_i(\\rv{c};\\rv{c}) &= \\nabla h_0(\\rv{c}) + \\sum_{i=1}^m \\mu_i\\nabla h_i(\\rv{c}),\\\\\n\\lb{g}_i(\\rv{c};\\rv{c}) &\\ge h_i(\\rv{c}) &\\forall i=1,2,\\ldots,m,\\\\\n\\mu_i &\\ge 0 &\\forall i=1,2,\\ldots,m,\\\\\n\\mu_i\\lb{g}_i(\\rv{c};\\rv{c}) &= \\mu_i h_i(\\rv{c}) &\\forall i=1,2,\\ldots,m.\n\\end{aligned}\n\\]\nSince $\\lb{g}_i (\\rv{c};\\rv{c}) = g_i(\\rv{c})$ and $\\nabla \\lb{g}_i (\\rv{c};\\rv{c}) = \\nabla g_i(\\rv{c})$ for all $i=0,1,\\ldots,m$, we conclude that $\\rv{c}$ is a critical point of~\\eqref{eq:dco}.\n\\end{proof}\n\nWe propose to solve the DCO problem~\\eqref{eq:dco} with a sequential convex optimization approach given in Algorithm~\\ref{algo:ccp}, also known as concave-convex procedure. A proof of showing that a sequence $\\{\\rv{z}_k\\}_{k=0}^\\infty$ generated by Algorithm~\\ref{algo:ccp} converges to a critical point $\\rv{z}^*$ of the original DCO problem~\\eqref{eq:dco} can be found in~\\cite{marks1978,lanckriet2009}.\n\n\\begin{algorithm}\n\t\\caption{Sequential convex optimization}\n\t\\label{algo:ccp}\n\t\\begin{algorithmic}[1]\n\t\t\\STATE Initialization: choose a feasible solution $\\rv{z}_0$ and a stopping criteria $\\varepsilon > 0$.\n\t\t\\STATE $\\rv{z}_1 \\in \\arg\\max \\{\\lb{g}_0(\\rv{z};\\rv{z}_0) - h_0(\\rv{z})\\colon \\lb{g}_i(\\rv{z};\\rv{z}_0) - h_i(\\rv{z}) \\ge 0, i =1,2,\\ldots,m\\}$\n\t\t\\STATE $k := 1$\n\t\t\\WHILE{$\\frac{\\|\\rv{z}_k - \\rv{z}_{k-1}\\|}{\\|\\rv{z}_k\\|} > \\varepsilon$}\n\t\t\\STATE $\\rv{z}_{k+1} \\in \\arg\\max \\{\\lb{g}_0(\\rv{z};\\rv{z}_k) - h_0(\\rv{z})\\colon \\lb{g}_i(\\rv{z};\\rv{z}_k) - h_i(\\rv{z}) \\ge 0, i =1,2,\\ldots,m\\}$\n\t\t\\STATE $k := k+1$\n\t\t\\ENDWHILE\n\t\\end{algorithmic}\n\\end{algorithm}\n\n\\section{Computational results}\\label{sec:results}\nProblem~\\eqref{eq:ngon:area} was solved in MATLAB using CVX~2.2 with MOSEK~9.1.9 and \\texttt{default precision} (tolerance $\\epsilon = 1.49 \\times 10^{-8}$). All the computations were carried out on an \\texttt{Intel(R) Core(TM) i7-3540M CPU @ 3.00 GHz} computing platform. Algorithm~\\ref{algo:ccp} was implemented as a MATLAB package: OPTIGON, which is freely available at \\url{https://github.com/cbingane/optigon}. OPTIGON requires that CVX be installed. CVX is a MATLAB-based modeling system for convex\noptimization, which turns MATLAB into a modeling language, allowing constraints and objectives to be specified using standard MATLAB expression syntax~\\cite{cvx2}.\n\nWe chose the following values as initial solution:\n\\[\n\\begin{aligned}\na_0 &= 0, &b_0 &= 0,\\\\\na_i &= \\frac{\\sin \\frac{2i\\pi}{n-1}}{2\\cos \\frac{\\pi}{2n-2}} = -a_{n-i}, & b_i &= \\frac{1-\\cos \\frac{2i\\pi}{n-1}}{2\\cos \\frac{\\pi}{2n-2}} = b_{n-i} &\\forall i=1,\\ldots,n/2-1,\\\\\na_{n/2} &= 0, &b_{n/2} &= 1,\n\\end{aligned}\n\\]\nwhich define the $n$-gon $\\geo{R}_{n-1}^+$, and the stopping criteria $\\varepsilon = 10^{-5}$. Table~\\ref{table:area} shows the optimal values $A_n^*$ of the maximal area problem for even numbers $n=6,8,\\ldots,128$, along with the areas of the initial $n$-gons $\\geo{R}_{n-1}^+$, the best lower bounds $\\lb{A}_n$ found in the literature, and the upper bounds~$\\ub{A}_n$. We also report the number $k$ of iterations in Algotithm~\\ref{algo:ccp} for each~$n$. The results support the following keypoints:\n\\begin{enumerate}\n\\item For $6 \\le n\\le 12$, $\\lb{A}_n - A_n^* \\le 10^{-8}$, i.e., Algorithm~\\ref{algo:ccp} converges to the best known optimal solutions found in the literature.\n\\item For $32 \\le n \\le 80$, $\\lb{A}_n < A(\\geo{R}_{n-1}^+) < A_n^*$, i.e., the solutions obtained by Pinter~\\cite{pinter2018} are suboptimal.\n\\item For all $n$, the solutions obtained with Algorithm~\\ref{algo:ccp} verify, within the limit of the numerical computations, Theorem~\\ref{thm:area:diam} and Conjecture~\\ref{thm:area:sym}, i.e.,\n\\[\n\\begin{aligned}\nx_{n/2} &=0, &y_{n/2} &=1,\\\\\n\\|\\geo{v}_{n/2-1}\\| &=1, &\\|\\geo{v}_{n/2+1}\\| &=1,\\\\\n\\|\\geo{v}_{i+n/2}-\\geo{v}_i\\| &=1, &\\|\\geo{v}_{i+n/2+1}-\\geo{v}_i\\| &=1 &\\forall i=1,2,\\ldots,n/2-2,\\\\\n\\|\\geo{v}_{n-1}-\\geo{v}_{n/2-1}\\| &=1,\\\\\nx_{n-i} &=-x_i, &y_{n-i} &= y_i &\\forall i=1,2,\\ldots,n/2-1.\n\\end{aligned}\n\\]\nWe illustrate the largest small $16$-, $32$- and $64$-gons in Figure~\\ref{figure:Un}. Furthermore,  we remark that Theorem~\\ref{thm:area:diam} and Conjecture~\\ref{thm:area:sym} are verified by each polygon of the sequence generated by Algorithm~\\ref{algo:ccp}. All $6$-gons generated by the algorithm are represented in Figure~\\ref{figure:ccp:U6} and the coordinates of their vertices are given in Table~\\ref{table:ccp:U6}.\n\\end{enumerate}\n\n\n\\begin{table}[t]\n\t\\footnotesize\n\t\\centering\n\t\\caption{Maximal area problem}\n\t\\label{table:area}\n\t\\resizebox{!}{10cm}{\n\t\t\\begin{tabular}{@{}rlll|lr@{}}\n\t\t\t\\toprule\n\t\t\t$n$ & $A(\\geo{R}_{n-1}^+)$ & $\\lb{A}_n$ & $\\ub{A}_n$ & $A_n^*$ & \\# ite. $k$\t\\\\\n\t\t\t\\midrule\n\t\t\t6\t&\t0.6722882584\t&\t0.6749814429~\\cite{bieri1961,graham1975,mossinghoff2006b}\t&\t0.6961524227\t&\t0.6749814387\t&\t5\t\\\\\n\t\t\t8\t&\t0.7253199909\t&\t0.7268684828~\\cite{audet2002,mossinghoff2006b}\t&\t0.7350842599\t&\t0.7268684802\t&\t10\t\\\\\n\t\t\t10\t&\t0.7482573378\t&\t0.7491373459~\\cite{henrion2013,mossinghoff2006b}\t&\t0.7531627703\t&\t0.7491373454\t&\t16\t\\\\\n\t\t\t12\t&\t0.7601970055\t&\t0.7607298734~\\cite{henrion2013,mossinghoff2006b}\t&\t0.7629992851\t&\t0.7607298710\t&\t24\t\\\\\n\t\t\t14\t&\t0.7671877750\t&\t0.7675310111~\\cite{mossinghoff2006b}\t&\t0.7689359584\t&\t0.7675310093\t&\t33\t\\\\\n\t\t\t16\t&\t0.7716285345\t&\t0.7718613220~\\cite{mossinghoff2006b}\t&\t0.7727913493\t&\t0.7718613187\t&\t43\t\\\\\n\t\t\t18\t&\t0.7746235089\t&\t0.7747881651~\\cite{mossinghoff2006b}\t&\t0.7754356273\t&\t0.7747881619\t&\t55\t\\\\\n\t\t\t20\t&\t0.7767382147\t&\t0.7768587560~\\cite{mossinghoff2006b}\t&\t0.7773275822\t&\t0.7768587517\t&\t68\t\\\\\n\t\t\t22\t&\t0.7782865351\t&\t0.7783773308~\\cite{pinter2018}\t&\t0.7787276939\t&\t0.7783773228\t&\t81\t\\\\\n\t\t\t24\t&\t0.7794540033\t&\t0.7795240461~\\cite{pinter2018}\t&\t0.7797927529\t&\t0.7795240330\t&\t95\t\\\\\n\t\t\t26\t&\t0.7803559816\t&\t0.7804111201~\\cite{pinter2018}\t&\t0.7806217145\t&\t0.7804111058\t&\t109\t\\\\\n\t\t\t28\t&\t0.7810672517\t&\t0.7811114192~\\cite{pinter2018}\t&\t0.7812795297\t&\t0.7811114002\t&\t122\t\\\\\n\t\t\t30\t&\t0.7816380102\t&\t0.7816739255~\\cite{pinter2018}\t&\t0.7818102598\t&\t0.7816739044\t&\t136\t\\\\\n\t\t\t32\t&\t0.7821029651\t&\t0.7818946320~\\cite{pinter2018}\t&\t0.7822446490\t&\t0.7821325276\t&\t148\t\\\\\n\t\t\t34\t&\t0.7824867354\t&\t0.7823103007~\\cite{pinter2018}\t&\t0.7826046775\t&\t0.7825113660\t&\t159\t\\\\\n\t\t\t36\t&\t0.7828071755\t&\t0.7826513767~\\cite{pinter2018}\t&\t0.7829063971\t&\t0.7828279054\t&\t169\t\\\\\n\t\t\t38\t&\t0.7830774889\t&\t0.7829526627~\\cite{pinter2018}\t&\t0.7831617511\t&\t0.7830950955\t&\t177\t\\\\\n\t\t\t40\t&\t0.7833076096\t&\t0.7832011589~\\cite{pinter2018}\t&\t0.7833797744\t&\t0.7833226804\t&\t183\t\\\\\n\t\t\t42\t&\t0.7835051276\t&\t0.7834135187~\\cite{pinter2018}\t&\t0.7835674041\t&\t0.7835181187\t&\t185\t\\\\\n\t\t\t44\t&\t0.7836759223\t&\t0.7835966860~\\cite{pinter2018}\t&\t0.7837300377\t&\t0.7836871900\t&\t184\t\\\\\n\t\t\t46\t&\t0.7838246055\t&\t0.7837554636~\\cite{pinter2018}\t&\t0.7838719255\t&\t0.7838344336\t&\t179\t\\\\\n\t\t\t48\t&\t0.7839548353\t&\t0.7838942710~\\cite{pinter2018}\t&\t0.7839964516\t&\t0.7839634510\t&\t172\t\\\\\n\t\t\t50\t&\t0.7840695435\t&\t0.7840161496~\\cite{pinter2018}\t&\t0.7841063371\t&\t0.7840771278\t&\t162\t\\\\\n\t\t\t52\t&\t0.7841711020\t&\t0.7841233641~\\cite{pinter2018}\t&\t0.7842037903\t&\t0.7841778072\t&\t150\t\\\\\n\t\t\t54\t&\t0.7842614465\t&\t0.7842192995~\\cite{pinter2018}\t&\t0.7842906181\t&\t0.7842674010\t&\t138\t\\\\\n\t\t\t56\t&\t0.7843421691\t&\t0.7843044654~\\cite{pinter2018}\t&\t0.7843683109\t&\t0.7843474779\t&\t128\t\\\\\n\t\t\t58\t&\t0.7844145892\t&\t0.7843807534~\\cite{pinter2018}\t&\t0.7844381066\t&\t0.7844193386\t&\t118\t\\\\\n\t\t\t60\t&\t0.7844798073\t&\t0.7844492943~\\cite{pinter2018}\t&\t0.7845010402\t&\t0.7844840717\t&\t109\t\\\\\n\t\t\t62\t&\t0.7845387477\t&\t0.7845111362~\\cite{pinter2018}\t&\t0.7845579827\t&\t0.7845425886\t&\t101\t\\\\\n\t\t\t64\t&\t0.7845921910\t&\t0.7834620877~\\cite{pinter2018}\t&\t0.7846096710\t&\t0.7845956631\t&\t94\t\\\\\n\t\t\t66\t&\t0.7846408000\t&\t0.7845910589~\\cite{pinter2018}\t&\t0.7846567322\t&\t0.7846439473\t&\t88\t\\\\\n\t\t\t68\t&\t0.7846851407\t&\t0.7846139029~\\cite{pinter2018}\t&\t0.7846997026\t&\t0.7846880001\t&\t82\t\\\\\n\t\t\t70\t&\t0.7847256986\t&\t0.7846403575~\\cite{pinter2018}\t&\t0.7847390429\t&\t0.7847283036\t&\t77\t\\\\\n\t\t\t72\t&\t0.7847628920\t&\t0.7847454020~\\cite{pinter2018}\t&\t0.7847751508\t&\t0.7847652718\t&\t72\t\\\\\n\t\t\t74\t&\t0.7847970830\t&\t0.7845564840~\\cite{pinter2018}\t&\t0.7848083708\t&\t0.7847992622\t&\t68\t\\\\\n\t\t\t76\t&\t0.7848285863\t&\t0.7847585719~\\cite{pinter2018}\t&\t0.7848390031\t&\t0.7848305850\t&\t64\t\\\\\n\t\t\t78\t&\t0.7848576763\t&\t0.7845160579~\\cite{pinter2018}\t&\t0.7848673094\t&\t0.7848595143\t&\t61\t\\\\\n\t\t\t80\t&\t0.7848845934\t&\t0.7848252941~\\cite{pinter2018}\t&\t0.7848935195\t&\t0.7848862871\t&\t58\t\\\\\n\t\t\t82\t&\t0.7849095487\t&\t--\t&\t0.7849178354\t&\t0.7849111119\t&\t55\t\\\\\n\t\t\t84\t&\t0.7849327284\t&\t--\t&\t0.7849404352\t&\t0.7849341725\t&\t52\t\\\\\n\t\t\t86\t&\t0.7849542969\t&\t--\t&\t0.7849614768\t&\t0.7849556352\t&\t50\t\\\\\n\t\t\t88\t&\t0.7849744002\t&\t--\t&\t0.7849811001\t&\t0.7849756425\t&\t48\t\\\\\n\t\t\t90\t&\t0.7849931681\t&\t--\t&\t0.7849994298\t&\t0.7849943223\t&\t46\t\\\\\n\t\t\t92\t&\t0.7850107163\t&\t--\t&\t0.7850165772\t&\t0.7850117894\t&\t44\t\\\\\n\t\t\t94\t&\t0.7850271482\t&\t--\t&\t0.7850326419\t&\t0.7850281477\t&\t42\t\\\\\n\t\t\t96\t&\t0.7850425565\t&\t--\t&\t0.7850477130\t&\t0.7850434878\t&\t40\t\\\\\n\t\t\t98\t&\t0.7850570245\t&\t--\t&\t0.7850618708\t&\t0.7850578951\t&\t39\t\\\\\n\t\t\t100\t&\t0.7850706272\t&\t--\t&\t0.7850751877\t&\t0.7850714422\t&\t38\t\\\\\n\t\t\t102\t&\t0.7850834323\t&\t--\t&\t0.7850877290\t&\t0.7850841941\t&\t36\t\\\\\n\t\t\t104\t&\t0.7850955008\t&\t--\t&\t0.7850995538\t&\t0.7850962152\t&\t35\t\\\\\n\t\t\t106\t&\t0.7851068883\t&\t--\t&\t0.7851107156\t&\t0.7851075587\t&\t34\t\\\\\n\t\t\t108\t&\t0.7851176450\t&\t--\t&\t0.7851212630\t&\t0.7851182747\t&\t33\t\\\\\n\t\t\t110\t&\t0.7851278167\t&\t--\t&\t0.7851312404\t&\t0.7851284086\t&\t32\t\\\\\n\t\t\t112\t&\t0.7851374450\t&\t--\t&\t0.7851406881\t&\t0.7851380017\t&\t31\t\\\\\n\t\t\t114\t&\t0.7851465680\t&\t--\t&\t0.7851496430\t&\t0.7851470916\t&\t30\t\\\\\n\t\t\t116\t&\t0.7851552203\t&\t--\t&\t0.7851581386\t&\t0.7851557129\t&\t29\t\\\\\n\t\t\t118\t&\t0.7851634339\t&\t--\t&\t0.7851662060\t&\t0.7851639010\t&\t29\t\\\\\n\t\t\t120\t&\t0.7851712379\t&\t--\t&\t0.7851738734\t&\t0.7851716781\t&\t28\t\\\\\n\t\t\t122\t&\t0.7851786591\t&\t--\t&\t0.7851811668\t&\t0.7851790741\t&\t27\t\\\\\n\t\t\t124\t&\t0.7851857221\t&\t--\t&\t0.7851881101\t&\t0.7851861129\t&\t26\t\\\\\n\t\t\t126\t&\t0.7851924497\t&\t--\t&\t0.7851947255\t&\t0.7851928211\t&\t26\t\\\\\n\t\t\t128\t&\t0.7851988626\t&\t--\t&\t0.7852010332\t&\t0.7851992126\t&\t25\t\\\\\n\t\t\t\\bottomrule\n\t\t\\end{tabular}\n\t}\n\\end{table}\n\n\\begin{figure}\n\t\\centering\n\t\\subfloat[$(\\geo{U}_{16},0.771861)$]{\n\t\\begin{tikzpicture}[scale=5]\n\t\\draw[dashed] (0,0) -- (0.2163,0.0539) -- (0.3801,0.1794) -- (0.4793,0.3573) -- (0.5000,0.5595) -- (0.4390,0.7532) -- (0.3070,0.9060) -- (0.1320,0.9912) -- (0,1) -- (-0.1320,0.9912) -- (-0.3070,0.9060) -- (-0.4390,0.7532) -- (-0.5000,0.5595) -- (-0.4793,0.3573) -- (-0.3801,0.1794) -- (-0.2163,0.0539) -- cycle;\n\t\\draw (0,0)--(0,1);\n\t\\draw (0,0)--(0.1320,0.9912);\\draw (0,0)--(-0.1320,0.9912);\n\t\\draw (0.2163,0.0539)--(-0.1320,0.9912);\\draw (0.2163,0.0539)--(-0.3070,0.9060);\n\t\\draw (0.3801,0.1794)--(-0.3070,0.9060);\\draw (0.3801,0.1794)--(-0.4390,0.7532);\n\t\\draw (0.4793,0.3573)--(-0.4390,0.7532);\\draw (0.4793,0.3573)--(-0.5000,0.5595);\n\t\\draw (0.5000,0.5595)--(-0.5000,0.5595);\\draw (0.5000,0.5595)--(-0.4793,0.3573);\n\t\\draw (0.4390,0.7532)--(-0.4793,0.3573);\\draw (0.4390,0.7532)--(-0.3801,0.1794);\n\t\\draw (0.3070,0.9060)--(-0.3801,0.1794);\\draw (0.3070,0.9060)--(-0.2163,0.0539);\n\t\\draw (0.1320,0.9912)--(-0.2163,0.0539);\n\t\\end{tikzpicture}\n}\n\\subfloat[$(\\geo{U}_{32},0.782133)$]{\n\t\\begin{tikzpicture}[scale=5]\n\t\\draw[dashed] (0,0) -- (0.1083,0.0131) -- (0.2043,0.0450) -- (0.2910,0.0947) -- (0.3661,0.1606) -- (0.4266,0.2401) -- (0.4702,0.3301) -- (0.4950,0.4271) -- (0.5000,0.5271) -- (0.4850,0.6261) -- (0.4507,0.7200) -- (0.3984,0.8052) -- (0.3302,0.8783) -- (0.2491,0.9363) -- (0.1587,0.9768) -- (0.0654,0.9979) -- (0,1) -- (-0.0654,0.9979) -- (-0.1587,0.9768) -- (-0.2491,0.9363) -- (-0.3302,0.8783) -- (-0.3984,0.8052) -- (-0.4507,0.7200) -- (-0.4850,0.6261) -- (-0.5000,0.5271) -- (-0.4950,0.4271) -- (-0.4702,0.3301) -- (-0.4266,0.2401) -- (-0.3661,0.1606) -- (-0.2910,0.0947) -- (-0.2043,0.0450) -- (-0.1083,0.0131) -- cycle;\n\t\\draw (0,0)--(0,1);\n\t\\draw (0,0)--(0.0654,0.9979);\\draw (0,0)--(-0.0654,0.9979);\n\t\\draw (0.1083,0.0131)--(-0.0654,0.9979);\\draw (0.1083,0.0131)--(-0.1587,0.9768);\n\t\\draw (0.2043,0.0450)--(-0.1587,0.9768);\\draw (0.2043,0.0450)--(-0.2491,0.9363);\n\t\\draw (0.2910,0.0947)--(-0.2491,0.9363);\\draw (0.2910,0.0947)--(-0.3302,0.8783);\n\t\\draw (0.3661,0.1606)--(-0.3302,0.8783);\\draw (0.3661,0.1606)--(-0.3984,0.8052);\n\t\\draw (0.4266,0.2401)--(-0.3984,0.8052);\\draw (0.4266,0.2401)--(-0.4507,0.7200);\n\t\\draw (0.4702,0.3301)--(-0.4507,0.7200);\\draw (0.4702,0.3301)--(-0.4850,0.6261);\n\t\\draw (0.4950,0.4271)--(-0.4850,0.6261);\\draw (0.4950,0.4271)--(-0.5000,0.5271);\n\t\\draw (0.5000,0.5271)--(-0.5000,0.5271);\\draw (0.5000,0.5271)--(-0.4950,0.4271);\n\t\\draw (0.4850,0.6261)--(-0.4950,0.4271);\\draw (0.4850,0.6261)--(-0.4702,0.3301);\n\t\\draw (0.4507,0.7200)--(-0.4702,0.3301);\\draw (0.4507,0.7200)--(-0.4266,0.2401);\n\t\\draw (0.3984,0.8052)--(-0.4266,0.2401);\\draw (0.3984,0.8052)--(-0.3661,0.1606);\n\t\\draw (0.3302,0.8783)--(-0.3661,0.1606);\\draw (0.3302,0.8783)--(-0.2910,0.0947);\n\t\\draw (0.2491,0.9363)--(-0.2910,0.0947);\\draw (0.2491,0.9363)--(-0.2043,0.0450);\n\t\\draw (0.1587,0.9768)--(-0.2043,0.0450);\\draw (0.1587,0.9768)--(-0.1083,0.0131);\n\t\\draw (0.0654,0.9979)--(-0.1083,0.0131);\n\t\\end{tikzpicture}\n}\n\\subfloat[$(\\geo{U}_{64},0.784596)$]{\n\t\\begin{tikzpicture}[scale=5]\n\t\\draw[dashed] (0,0) -- (0.0531,0.0031) -- (0.1018,0.0108) -- (0.1492,0.0231) -- (0.1953,0.0400) -- (0.2398,0.0615) -- (0.2820,0.0874) -- (0.3216,0.1174) -- (0.3581,0.1513) -- (0.3911,0.1887) -- (0.4201,0.2291) -- (0.4451,0.2723) -- (0.4656,0.3178) -- (0.4815,0.3650) -- (0.4926,0.4136) -- (0.4988,0.4631) -- (0.5000,0.5129) -- (0.4963,0.5627) -- (0.4876,0.6118) -- (0.4741,0.6598) -- (0.4559,0.7062) -- (0.4332,0.7506) -- (0.4061,0.7924) -- (0.3750,0.8314) -- (0.3403,0.8670) -- (0.3022,0.8990) -- (0.2612,0.9270) -- (0.2178,0.9507) -- (0.1725,0.9700) -- (0.1257,0.9846) -- (0.0783,0.9944) -- (0.0319,0.9995) -- (0,1) -- (-0.0319,0.9995) -- (-0.0783,0.9944) -- (-0.1257,0.9846) -- (-0.1725,0.9700) -- (-0.2178,0.9507) -- (-0.2612,0.9270) -- (-0.3022,0.8990) -- (-0.3403,0.8670) -- (-0.3750,0.8314) -- (-0.4061,0.7924) -- (-0.4332,0.7506) -- (-0.4559,0.7062) -- (-0.4741,0.6598) -- (-0.4876,0.6118) -- (-0.4963,0.5627) -- (-0.5000,0.5129) -- (-0.4988,0.4631) -- (-0.4926,0.4136) -- (-0.4815,0.3650) -- (-0.4656,0.3178) -- (-0.4451,0.2723) -- (-0.4201,0.2291) -- (-0.3911,0.1887) -- (-0.3581,0.1513) -- (-0.3216,0.1174) -- (-0.2820,0.0874) -- (-0.2398,0.0615) -- (-0.1953,0.0400) -- (-0.1492,0.0231) -- (-0.1018,0.0108) -- (-0.0531,0.0031) -- cycle;\n\t\\draw (0,0)--(0,1);\n\t\\draw (0,0)--(0.0319,0.9995);\\draw (0,0)--(-0.0319,0.9995);\n\t\\draw (0.0531,0.0031)--(-0.0319,0.9995);\\draw (0.0531,0.0031)--(-0.0783,0.9944);\n\t\\draw (0.1018,0.0108)--(-0.0783,0.9944);\\draw (0.1018,0.0108)--(-0.1257,0.9846);\n\t\\draw (0.1492,0.0231)--(-0.1257,0.9846);\\draw (0.1492,0.0231)--(-0.1725,0.9700);\n\t\\draw (0.1953,0.0400)--(-0.1725,0.9700);\\draw (0.1953,0.0400)--(-0.2178,0.9507);\n\t\\draw (0.2398,0.0615)--(-0.2178,0.9507);\\draw (0.2398,0.0615)--(-0.2612,0.9270);\n\t\\draw (0.2820,0.0874)--(-0.2612,0.9270);\\draw (0.2820,0.0874)--(-0.3022,0.8990);\n\t\\draw (0.3216,0.1174)--(-0.3022,0.8990);\\draw (0.3216,0.1174)--(-0.3403,0.8670);\n\t\\draw (0.3581,0.1513)--(-0.3403,0.8670);\\draw (0.3581,0.1513)--(-0.3750,0.8314);\n\t\\draw (0.3911,0.1887)--(-0.3750,0.8314);\\draw (0.3911,0.1887)--(-0.4061,0.7924);\n\t\\draw (0.4201,0.2291)--(-0.4061,0.7924);\\draw (0.4201,0.2291)--(-0.4332,0.7506);\n\t\\draw (0.4451,0.2723)--(-0.4332,0.7506);\\draw (0.4451,0.2723)--(-0.4559,0.7062);\n\t\\draw (0.4656,0.3178)--(-0.4559,0.7062);\\draw (0.4656,0.3178)--(-0.4741,0.6598);\n\t\\draw (0.4815,0.3650)--(-0.4741,0.6598);\\draw (0.4815,0.3650)--(-0.4876,0.6118);\n\t\\draw (0.4926,0.4136)--(-0.4876,0.6118);\\draw (0.4926,0.4136)--(-0.4963,0.5627);\n\t\\draw (0.4988,0.4631)--(-0.4963,0.5627);\\draw (0.4988,0.4631)--(-0.5000,0.5129);\n\t\\draw (0.5000,0.5129)--(-0.5000,0.5129);\\draw (0.5000,0.5129)--(-0.4988,0.4631);\n\t\\draw (0.4963,0.5627)--(-0.4988,0.4631);\\draw (0.4963,0.5627)--(-0.4926,0.4136);\n\t\\draw (0.4876,0.6118)--(-0.4926,0.4136);\\draw (0.4876,0.6118)--(-0.4815,0.3650);\n\t\\draw (0.4741,0.6598)--(-0.4815,0.3650);\\draw (0.4741,0.6598)--(-0.4656,0.3178);\n\t\\draw (0.4559,0.7062)--(-0.4656,0.3178);\\draw (0.4559,0.7062)--(-0.4451,0.2723);\n\t\\draw (0.4332,0.7506)--(-0.4451,0.2723);\\draw (0.4332,0.7506)--(-0.4201,0.2291);\n\t\\draw (0.4061,0.7924)--(-0.4201,0.2291);\\draw (0.4061,0.7924)--(-0.3911,0.1887);\n\t\\draw (0.3750,0.8314)--(-0.3911,0.1887);\\draw (0.3750,0.8314)--(-0.3581,0.1513);\n\t\\draw (0.3403,0.8670)--(-0.3581,0.1513);\\draw (0.3403,0.8670)--(-0.3216,0.1174);\n\t\\draw (0.3022,0.8990)--(-0.3216,0.1174);\\draw (0.3022,0.8990)--(-0.2820,0.0874);\n\t\\draw (0.2612,0.9270)--(-0.2820,0.0874);\\draw (0.2612,0.9270)--(-0.2398,0.0615);\n\t\\draw (0.2178,0.9507)--(-0.2398,0.0615);\\draw (0.2178,0.9507)--(-0.1953,0.0400);\n\t\\draw (0.1725,0.9700)--(-0.1953,0.0400);\\draw (0.1725,0.9700)--(-0.1492,0.0231);\n\t\\draw (0.1257,0.9846)--(-0.1492,0.0231);\\draw (0.1257,0.9846)--(-0.1018,0.0108);\n\t\\draw (0.0783,0.9944)--(-0.1018,0.0108);\\draw (0.0783,0.9944)--(-0.0531,0.0031);\n\t\\draw (0.0319,0.9995)--(-0.0531,0.0031);\n\t\\end{tikzpicture}\n}\n\\caption{Three largest small $n$-gons $(\\geo{U}_n,A_n^*)$}\n\\label{figure:Un}\n\\end{figure}\n\n\\begin{figure}[h]\n\t\\centering\n\t\\subfloat[$(\\geo{R}_5^+,0.672288)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\t\\draw[dashed] (0,0) -- (0.5000,0.3633) -- (0.3090,0.9511) -- (0,1) -- (-0.3090,0.9511) -- (-0.5000,0.3633) -- cycle;\n\t\t\t\\draw (0,1) -- (0,0) -- (0.3090,0.9511) -- (-0.5000,0.3633) -- (0.5000,0.3633) -- (-0.3090,0.9511) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\\subfloat[$(\\geo{P}_6^1,0.674941)$]{\n\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.5000,0.3975) -- (0.3397,0.9405) -- (0,1) -- (-0.3397,0.9405) -- (-0.5000,0.3975) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.3397,0.9405) -- (-0.5000,0.3975) -- (0.5000,0.3975) -- (-0.3397,0.9405) -- (0,0);\n\t\\end{tikzpicture}\n}\n\\subfloat[$(\\geo{P}_6^2,0.674981)$]{\n\t\\begin{tikzpicture}[scale=4]\n\t\t\\draw[dashed] (0,0) -- (0.5000,0.4018) -- (0.3433,0.9392) -- (0,1) -- (-0.3433,0.9392) -- (-0.5000,0.4024) -- cycle;\n\t\t\\draw (0,1) -- (0,0) -- (0.3433,0.9392) -- (-0.5000,0.4018) -- (0.5000,0.4018) -- (-0.3433,0.9392) -- (0,0);\n\t\\end{tikzpicture}\n}\\\\\n\t\\subfloat[$(\\geo{P}_6^3,0.674981)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\t\\draw[dashed] (0,0) -- (0.5000,0.4023) -- (0.3437,0.9391) -- (0,1) -- (-0.3437,0.9391) -- (-0.5000,0.4023) -- cycle;\n\t\t\t\\draw (0,1) -- (0,0) -- (0.3437,0.9391) -- (-0.5000,0.4023) -- (0.5000,0.4023) -- (-0.3437,0.9391) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{P}_6^4,0.674981)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\t\\draw[dashed] (0,0) -- (0.5000,0.4023) -- (0.3438,0.9391) -- (0,1) -- (-0.3438,0.9391) -- (-0.5000,0.4023) -- cycle;\n\t\t\t\\draw (0,1) -- (0,0) -- (0.3438,0.9391) -- (-0.5000,0.4023) -- (0.5000,0.4023) -- (-0.3438,0.9391) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\subfloat[$(\\geo{P}_6^5,0.674981)$]{\n\t\t\\begin{tikzpicture}[scale=4]\n\t\t\t\\draw[dashed] (0,0) -- (0.5000,0.4024) -- (0.3438,0.9391) -- (0,1) -- (-0.3438,0.9391) -- (-0.5000,0.4024) -- cycle;\n\t\t\t\\draw (0,1) -- (0,0) -- (0.3438,0.9391) -- (-0.5000,0.4024) -- (0.5000,0.4024) -- (-0.3438,0.9391) -- (0,0);\n\t\t\\end{tikzpicture}\n\t}\n\t\\caption{All $6$-gons $(\\geo{P}_6^k,A(\\geo{P}_6^k))$ generated by Algorithm~\\ref{algo:ccp}}\n\t\\label{figure:ccp:U6}\n\\end{figure}\n\n\\begin{table}[t]\n\t\\footnotesize\n\t\\centering\n\t\\caption{Vertices of $6$-gons generated by Algorithm~\\ref{algo:ccp}}\n\t\\label{table:ccp:U6}\n\t\t\\resizebox{\\linewidth}{!}{\n\t\\begin{tabular}{@{}l|ccccc|l@{}}\n\t\t\\toprule\n\t\t$6$-gon & \\multicolumn{5}{c|}{Coordinates $(x_i,y_i)$} & Area \\\\\n\t\t\\cmidrule{2-6} & $(x_1,y_1)$ & $(x_2,y_2)$ & $(x_3,y_3)$ &\t$(x_4,y_4)$ & $(x_5,y_5)$\t&\\\\\n\t\t\\midrule\n\t\t$\\geo{R}_5^+$ & $(0.500000,0.363271)$ & $(0.309017,0.951057)$ & $(0.000000,1.000000)$ &\t$(-0.309017,0.951057)$ & $(-0.500000,0.363271)$\t&\t$0.6722882584$\t\\\\\n\t\t$\\geo{P}_6^1$ & $(0.500000,0.397460)$ & $(0.339680,0.940541)$ & $(0.000000,1.000000)$ &\t$(-0.339680,0.940541)$ & $(-0.500000,0.397460)$\t&\t$0.6749414624$\t\\\\\n\t\t$\\geo{P}_6^2$ & $(0.500000,0.401764)$ & $(0.343285,0.939231)$ & $(0.000000,1.000000)$ &\t$(-0.343285,0.939231)$ & $(-0.500000,0.401764)$\t&\t$0.6749808685$\t\\\\\n\t\t$\\geo{P}_6^3$ & $(0.500000,0.402283)$ & $(0.343715,0.939074)$ & $(0.000000,1.000000)$ &\t$(-0.343715,0.939074)$ & $(-0.500000,0.402283)$\t&\t$0.6749814310$\t\\\\\n\t\t$\\geo{P}_6^4$ & $(0.500000,0.402345)$ & $(0.343766,0.939055)$ & $(0.000000,1.000000)$ &\t$(-0.343766,0.939055)$ & $(-0.500000,0.402345)$\t&\t$0.6749814386$\t\\\\\n\t\t$\\geo{P}_6^5$ & $(0.500000,0.402352)$ & $(0.343773,0.939053)$ & $(0.000000,1.000000)$ &\t$(-0.343773,0.939053)$ & $(-0.500000,0.402352)$\t&\t$0.6749814387$\t\\\\\n\t\t\\bottomrule\n\\end{tabular}\n\t}\n\\end{table}\n\n\\section{Conclusion}\\label{sec:conclusion}\nWe proposed a sequential convex optimization approach to find the largest small $n$-gon for a given even number $n\\ge 6$, which is formulated as a nonconvex quadratically constrained quadratic optimization problem. The algorithm, also known as the concave-convex procedure, guarantees convergence to a locally optimal solution.\n\nWithout assuming Graham's conjecture nor the existence of an axis of symmetry in our quadratic formulation, numerical experiments on polygons with up to $n=128$ sides showed that each optimal $n$-gon obtained with the algorithm proposed verifies both conditions within the limitation of the numerical computations. Futhermore, for even $6\\le n\\le 12$, the $n$-gons obtained correspond to the known largest small $n$-gons.\n\n\n\\section*{Acknowledgements}\nThe author thanks Charles Audet, Professor at Polytechnique Montreal, for helpful discussions on largest small polygons and helpful comments on early drafts of this paper.\n\n\\bibliographystyle{ieeetr}\n", "meta": {"timestamp": "2021-06-02T02:11:56", "yymm": "2009", "arxiv_id": "2009.07893", "language": "en", "url": "https://arxiv.org/abs/2009.07893", "abstract": "A small polygon is a polygon of unit diameter. The maximal area of a small polygon with $n=2m$ vertices is not known when $m\\ge 7$. Finding the largest small $n$-gon for a given number $n\\ge 3$ can be formulated as a nonconvex quadratically constrained quadratic optimization problem. We propose to solve this problem with a sequential convex optimization approach, which is an ascent algorithm guaranteeing convergence to a locally optimal solution. Numerical experiments on polygons with up to $n=128$ sides suggest that the optimal solutions obtained are near-global. Indeed, for even $6 \\le n \\le 12$, the algorithm proposed in this work converges to known global optimal solutions found in the literature.", "subjects": "Optimization and Control (math.OC)", "title": "Largest small polygons: A sequential convex optimization approach", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9890130599601177, "lm_q2_score": 0.8519528019683106, "lm_q1q2_score": 0.842592447616275}}
{"url": "https://arxiv.org/abs/2111.14076", "title": "Structural theorems on the distance sets over finite fields", "abstract": "Let $\\mathbb{F}_q$ be a finite field of order $q$. Iosevich and Rudnev (2005) proved that for any set $A\\subset \\mathbb{F}_q^d$, if $|A|\\gg q^{\\frac{d+1}{2}}$, then the distance set $\\Delta(A)$ contains a positive proportion of all distances. Although this result is sharp in odd dimensions, it is conjectured that the right exponent should be $\\frac{d}{2}$ in even dimensions. During the last 15 years, only some improvements have been made in two dimensions, and the conjecture is still wide open in higher dimensions. To fill the gap, we need to understand more about the structures of the distance sets, the main purpose of this paper is to provide some structural theorems on the distribution of square and non-square distances.", "text": "\\section{Introduction}\n\n\nLet $\\mathbb{F}_q$ be a finite field of order $q$ which is a prime power. The distance between two points $x$ and $y$ in the space $\\mathbb{F}_q^d$ is defined by \\[||x-y||:=(x_1-y_1)^2+\\cdots+(x_d-y_d)^2.\\]\nFor $A\\subset \\mathbb{F}_q^d$, let $\\Delta(A)$ denote the set of distances determined by pairs of points in $A$, i.e. \n\\[\\Delta(A)=\\{||x-y||\\colon x, y\\in A\\}.\\]\nStudying the magnitude of $\\Delta(E)$ has been received much attention since 2007 due to its connection to the Falconer distance conjecture in Geometric Measure Theory, which says that for any compact set $A\\subset \\mathbb{R}^d$ of Hausdorff dimension greater than $d/2$, the distance set $\\Delta(A)$ is of positive Lebesgue measure. Recent progress on this conjecture can be found in \\cite{Du1, Du3, Du2, alex-fal}.\n\nIn the finite field setting, Iosevich and Rudnev \\cite{IR07} proved that for $A\\subset \\mathbb{F}_q^d$, if $|A|\\ge 4q^{\\frac{d+1}{2}}$, then $\\Delta(A)=\\mathbb{F}_q$. The exponent $\\frac{d+1}{2}$ is sharp in odd dimensions, namely, $d\\equiv 3\\mod 4$ and $q\\equiv 1\\mod 4,$ or $d\\equiv 1\\mod 4$, we refer the reader to \\cite{HIKR10} for constructions. In even dimensions, it is conjectured that the right exponent should be $\\frac{d}{2}$ which is directly in line with the Falconer distance conjecture. In a very recent paper, Murphy, Petridis, Pham, Rudnev, and Stevens \\cite{MPT} obtained the exponent $\\frac{5}{4}$ in the plane over prime fields, which improves the exponent $\\frac{4}{3}$ by Chapman, Erdogan, Hart, Iosevich, and Koh in \\cite{CEHIK10}. In particular, they showed that for $A\\subset \\mathbb{F}_p^2$ with $|A|\\gg p^{5/4}$, then there exists $x\\in A$ such that\n\\begin{equation}\\label{54}|\\Delta_x(A)|\\gg p,\\end{equation}\nwhere $\\Delta_x(A):=\\{||x-a||\\colon a\\in A\\}$. This matches the breakthrough dimensional threshold $\\frac{5}{4}$ on the Falconer distance conjecture given by Guth, Iosevich, Ou, and Wang in \\cite{alex-fal}. For $4p<|A|\\le p^{5/4}$, Murphy et al. also proved that there exists $x\\in A$ with\n\\begin{equation}\\label{dim}\n    |\\Delta_x(A)|\\gg \\frac{|A|^{4/3}}{p^{2/3}}.\n\\end{equation}\nThis statement is similar to a result of Liu in \\cite{liu} for the continuous setting which states that for any compact set $A\\subset \\mathbb{R}^2$, if the Hausdorff dimension of $A$ is greater than $1$, then \n\\[\\dim_H(\\Delta_x(A))\\ge \\min \\left\\{\\frac{4}{3}\\dim_H(A)-\\frac{2}{3}, ~1 \\right\\},\\]\nfor some $x\\in A$. We note that when the dimension of $A$ is very close to $1$, then Shmerkin \\cite{hai} obtained better results, namely, \n\\[\\dim_H(\\Delta_x(A))\\ge \\frac{29}{42},\\]\nand \n\\[\\dim_H(\\Delta(A))\\ge \\frac{40}{57}.\\]\nOver finite fields, in higher even dimensions or $d\\equiv 3\\mod 4$ and $q\\equiv 3\\mod 4$, the best current exponent is still $\\frac{d+1}{2}$ due to Iosevich and Rudnev \\cite{IR07}, which is far from the conjecture $\\frac{d}{2}$. In order make further progress, we need to understand more about structures of the distance sets, and in this paper, we focus on the distribution of pairs of points of square and non-square distances in a given set $A\\subset \\mathbb{F}_q^d$.\n\nTo state our main results, we need the following definition.\n\\begin{definition} Let $\\eta$ denote the quadratic character of $\\mathbb F_q$ with $\\eta(0)=0.$ For $A\\subset \\mathbb F_q^d,$ we define \n$$  \\mathcal{SQ}(A):=|\\{(x, y)\\in A\\times A: \\eta(||x-y||)=1\\}|,$$\nand \n$$ \\mathcal{ZR}(A):=|\\{(x, y)\\in A\\times A: \\eta(||x-y||)=0\\}|,$$\n\nas the sets of pairs of square distances and zero-distance, respectively. \n\\end{definition}\n\nFor a given $A\\subset \\mathbb{F}_q^d$, if we consider the zero distance as a square, then either the number of pairs of square distances or the number of pairs of non-square distances is at least $|A|^2/2$. \n\nIn the first theorem, we show that when the size of $A$ is not small, most of the pairs of points in $A$ are of non-square distances. Precisely, we have  \n\\newpage\n\\begin{theorem}\\label{mainN0} Let $A\\subset \\mathbb F_q^d.$ \n \\begin{enumerate}\n \\item If $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n $$\\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q} \n - \\frac{|A|^2}{2 q^{\\frac{d-1}{2}}} - \\frac{|A|^2}{2 q^{\\frac{d+1}{2}}}+ q^{\\frac{d-1}{2}} |A|.$$\n \\item If $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n $$\\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q}+ \\frac{q^{\\frac{d+1}{2}} |A|}{2} -\\frac{q^{\\frac{d-1}{2}} |A|}{2}.$$\n \\item If $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n $$\\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q}-\\frac{|A|^2}{q^{\\frac{d}{2}}}+ \\frac{q^{\\frac{d}{2}} |A|}{2} +\\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n \\item If $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n $$ \\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q}+ \\frac{q^{\\frac{d}{2}} |A|}{2}-\\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n \\end{enumerate}\n \\end{theorem}\n \nWe say that a set $A\\subset \\mathbb F_q^d$ is a \\textit{square distance} set if for all $x,y\\in A, ~||x-y||$ is either zero or a square number of $\\mathbb F_q.$ It is clear that for such a set, one has $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)=|A|^2$ for any square distance set $A$ in $\\mathbb F_q^d.$ \nHence, if $A\\subset \\mathbb F_q^d$ is a square distance set, then $|A|^2$ is dominated by the upper bound of $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)$ given in Theorem \\ref{mainN0}. By solving the inequalities in terms of $|A|$,  we recover the following results of Iosevich, Shparlinski, and Xiong \\cite{iosevich}. We also refer the reader to \\cite{i1, i2, i3, iosevich} for more discussions and the motivation of this problem.\n \\begin{corollary}[Iosevich-Shparlinski-Xiong, \\cite{iosevich}]\\label{cor:1.3}\n Let $A$  be a square distance set in $\\mathbb F_q^d.$\n \\begin{enumerate}\n \\item If $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n $$|A| \\le \\frac{2 q^{\\frac{d+1}{2}}}{q-1+ (q+1) q^{-\\frac{(d-1)}{2}}}.$$ \n\n \\item If $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n $$ |A|\\le q^{\\frac{d+1}{2}}.$$\n \\item If $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n $$|A| \\le q^{\\frac{d}{2}}+ \\frac{2 (q^{\\frac{d}{2}}-q)}{q-1+ 2 q^{-\\frac{(d-2)}{2}}}.$$ \n\n \\item If $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n  $$ |A|\\le q^{\\frac{d}{2}}.$$\n \\end{enumerate}\n \\end{corollary}\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nIn the next two theorems, we are interested in the distribution of non-zero square distances in a given set $A\\subset \\mathbb{F}_q^d$.\n\\begin{theorem}[odd dimensions]\\label{mainNOdd} Let $A$ be a subset of $\\mathbb F_q^d.$ \n \\begin{enumerate} \n\\item Let $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4}.$ \n\n If $|A|\\ge (q^{(d+1)/2} +q)/( 1+ q^{-(d-1)/2}),$ then \n$$ \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + q^{\\frac{d-1}{2}} |A| -\\frac{|A|^2}{2q}- \\frac{|A|^2}{2 q^{\\frac{d-1}{2}}} - \\frac{|A|^2}{2 q^{\\frac{d+1}{2}}}.$$\nIf $|A|\\le (q^{(d+1)/2} +q)/( 1+ q^{-(d-1)/2}),$ then \n$$ \\mathcal{SQ}(A) \\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d-1}{2}} |A|}{2}\n  - \\frac{|A|^2}{2 q^{\\frac{d-1}{2}}} -\\frac{|A|}{2}.$$\n\n\n  \\item  Let $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then \n $$ \\mathcal{SQ}(A) \\le \\frac{|A|^2}{2} - \\frac{q^{\\frac{d-1}{2}} |A|}{2}\n  -\\frac{|A|}{2}+ \\min\\left\\{ \\frac{q^{\\frac{d+1}{2}} |A|}{2},~ \\frac{q^{\\frac{d-1}{2}} |A|}{2}+ \\frac{|A|^2}{2},~\\frac{|A|}{2}+ \\frac{q^{\\frac{d+1}{2}} |A|}{2}-\\frac{|A|^2}{2q} \\right\\}.$$\n \\end{enumerate}\n \\end{theorem}\n \\begin{theorem}[even dimensions] \\label{mainNEven} Let $A$ be a subset of $\\mathbb F_q^d.$ \n \n \\begin{enumerate}\n \\item\n If  $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n\n$$ \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}} |A|}{2} -\\frac{|A|^2}{2q}- \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n  \n\\item Let $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4}.$ \n\nIf $|A|\\ge (q^{d/2} +q)/( 1+ q^{-(d-2)/2}),$ then \n$$ \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}} |A|}{2}-\\frac{|A|^2}{q^{\\frac{d}{2}}} -\\frac{|A|^2}{2q}+ \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\nIf $|A|\\le (q^{d/2} +q)/( 1+q^{-(d-2)/2}),$ then\n$$ \\mathcal{SQ}(A) \\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}} |A|}{2}\n  - \\frac{|A|^2}{2 q^{\\frac{d}{2}}} -\\frac{|A|}{2}.$$\n\\end{enumerate}\n\\end{theorem}\nIt has been proved in \\cite{iosevich} that Corollary \\ref{cor:1.3} (2) and (4) are sharp, so the upper bounds of Theorem \\ref{mainN0} (2) and (4) are also best possible. It follows from Corollary \\ref{cor:1.3} that if the set of non-zero distances is fully contained in the group of squares, then the size of $A$ can not be bigger than a certain threshold. After posting this paper to Arxiv, Prof. Igor Shparlinski raised the question of studying the case for an arbitrary subgroup of $\\mathbb{F}_q\\setminus \\{0\\}$. We hope to address this question in a sequel paper. \n\n \n \n\n\n\n\\section{Discrete Fourier analysis and preliminary lemmas}\nIn this section, we recall notations from Discrete Fourier analysis and properties. \nLet $f$ be a complex valued function on $\\mathbb F_q^n,$ then its Fourier transform denoted by $\\widehat{f}$ is defined by\n$$ \\widehat{f}(m)=q^{-n} \\sum_{x\\in \\mathbb F_q^n} \\chi(-m\\cdot x) f(x),$$ \nhere $\\chi$ is the principal additive character of $\\mathbb F_q.$ \n\nWith this definition, the Fourier inversion theorem reads as\n$$ f(x)=\\sum_{m\\in \\mathbb F_q^n} \\chi(m\\cdot x) \\widehat{f}(m),$$\nwe also have the Plancherel theorem\n$$ \\sum_{m\\in \\mathbb F_q^n} |\\widehat{f}(m)|^2 =q^{-n}\\sum_{x\\in \\mathbb F_q^n} |f(x)|^2,$$\nwhich can be proved easily by using the following orthogonal property\n$$ \\sum_{\\alpha\\in \\mathbb F_q^n} \\chi(\\beta\\cdot \\alpha)\n=\\left\\{\\begin{array}{ll} 0\\quad &\\mbox{if}\\quad \\beta\\ne (0,\\ldots, 0),\\\\\nq^n\\quad &\\mbox{if}\\quad \\beta=(0,\\ldots,0). \\end{array}\\right.$$\n\nIn this paper, the quadratic character of $\\mathbb F_q$ is denoted by $\\eta$ with a convention that $\\eta(0)=0.$ For $a\\in \\mathbb F_q^*,$ the Gauss sum $G_a$ is defined by\n$$G_a=\\sum_{s\\in \\mathbb F_q^*}\\eta(s) \\chi(as),$$\nwhich can be rewritten as\n$$ G_a=\\sum_{s\\in \\mathbb F_q} \\chi(as^2)=\\eta(a) G_1.$$\nIt is not hard to see that $|G_a|=\\sqrt{q}$. In the following lemma, we recall the explicit form of the Gauss sum $G_1$ from \\cite[Theorem 5.15]{LN97}.\n\\begin{lemma}\\label{ExplicitGauss}\nLet $\\mathbb F_q$ be a finite field with $ q= p^{\\ell},$ where $p$ is an odd prime and $\\ell \\in {\\mathbb N}.$\nThen we have\n$$G_1= \\left\\{\\begin{array}{ll}  {(-1)}^{\\ell-1} q^{\\frac{1}{2}} \\quad &\\mbox{if} \\quad p \\equiv 1 \\mod 4 \\\\\n{(-1)}^{\\ell-1} i^\\ell q^{\\frac{1}{2}} \\quad &\\mbox{if} \\quad p\\equiv 3 \\mod 4.\\end{array}\\right. $$\n\\end{lemma}\n\n\n\\begin{corollary} \\label{Corm}Let $n$ be a positive integer. \n\\begin{enumerate}\n\\item\nIf $n\\equiv 0 \\mod {4}$ and $q\\equiv 3 \\mod{4},$ then \n$\\eta(-1)G_1^n= -q^{\\frac{n}{2}}.$\n\\item If $n\\equiv 2 \\mod{4},$ or $n\\equiv 0 \\mod{4}$ and $q\\equiv 1 \\mod{4}$, then \n$\\eta(-1)G_1^n= q^{\\frac{n}{2}}.$\n\\item If $n\\equiv 2 \\mod{4}$ and $q\\equiv 3 \\mod{4}$, then \n$ G_1^n= -q^{\\frac{n}{2}}.$\n\\item\nIf $n\\equiv 0 \\mod {4},$ or $n\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then \n$G_1^n= q^{\\frac{n}{2}}.$\n\\end{enumerate}\n\\end{corollary} \n\n\n\n\nThe following formula, which can be proved by completing the square and using a change of variables, will be used in our computations below\n\\begin{equation}\\label{ComSqu}  \n \\sum_{s\\in \\mathbb F_q} \\chi(as^2+bs)= \\eta(a)G_1 \\chi\\left(\\frac{b^2}{-4a}\\right),\\end{equation}\nwhere $a\\in \\mathbb F_q^*$ and $b\\in \\mathbb F_q.$\n\nLet $P\\in \\mathbb F_q[x_1,\\ldots,x_n]$ be a polynomial, we define a variety $V_P$ as\n$$V_P:=\\{x\\in \\mathbb F_q^n: P(x)=0\\}.$$\nIn the next lemma, we count the number of pairs of points $(x, y)$ in a given set in $\\mathbb{F}_q^n$ such that $x-y\\in V_P$. \n\n\\begin{lemma}\\label{AlexF} Let $E$ be a set in $\\mathbb{F}_q^n$. Then the number of pairs $(x,y)\\in E\\times E$ such that $P(x-y)=0$ is equal to \n$$ q^{2n}\\sum_{m\\in \\mathbb F_q^{n}} \\widehat{V_P}(m)|\\widehat{E}(m)|^2.$$\n\\end{lemma}\n\\begin{proof}\nLet $N$ be the number of pairs $(x,y)$ such that $P(x-y)=0$, i.e.\n$$ N=\\sum_{x,y\\in E} V_P(x-y).$$\nBy identifying the variety $V_P$ with the indicator function $1_{V_P}$ and using the Fourier inversion formula for the function $V_P(x-y)$ one has\n$$ N=\\sum_{x,y\\in \\mathbb F_q^n} E(x)E(y) \\sum_{m\\in \\mathbb F_q^n} \\widehat{V_P}(m) \\chi(m\\cdot (x-y)).$$ \nThen the lemma follows directly from the orthogonal property. \n\\end{proof}\n\nAs we shall see  in the proofs of our main results, given $A\\subset \\mathbb F_q^d,$ we shall relate the value $\\mathcal{SQ}(A) + \\mathcal{ZR}(A)$ to the Fourier decay of the cone in $\\mathbb F_q^{d+1}.$\nFor a positive integer $n\\ge 2,$ we recall that the cone $C_n$ is defined as\n\\begin{equation}\\label{conDe}C_n:=\\{x\\in \\mathbb F_q^n: x_n^2=x_1^2+x_2^2+\\cdots+ x_{n-1}^2\\}.\\end{equation}\n\n\\begin{definition}\nLet $x=(x_1,x_2,\\ldots, x_n) \\in \\mathbb F_q^n.$ We define \n$$ ||x||_{C_n}= x_1^2+x_2^2+\\cdots+ x_{n-1}^2 - x_n^2.$$\n\\end{definition}\nUsing the notation $|| \\cdot ||_{C_n}$,  the cone $C_n$ can be written by\n$$ C_n=\\{x\\in \\mathbb F_q^n: ||x||_{C_n}=0\\}.$$\nThe next lemma contains the Fourier transform formula for $C_n$ which was computed explicitly in \\cite{IMRN, KPP}.\n\\begin{lemma}\\label{FTForm} Let $C_n$ be the cone in $\\mathbb F_q^n$ defined in \\eqref{conDe}.\nThen, for any $m\\in \\mathbb F_q^n,$ we have\n\\begin{equation}\\label{FFC}\\widehat{C_n}(m)=q^{-1}\\delta_{0}(m) +q^{-n-1} \\eta(-1) G_1^{n} \\sum_{s\\ne 0} \\eta^n(s) \\chi\\left( \\frac{||m||_{C_n}}{-4s}\\right).\\end{equation}\n \\end{lemma}\n \n\n \n\n \n \n \\section{Preliminary settings on estimates of $\\mathcal{SQ}(A)$ and $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)$} \n For efficient estimates, we decompose $\\mathcal{SQ}(A)$ as the sum of $\\mathcal{SQ}(A) + \\mathcal{ZR}(A)/2$ and $-\\mathcal{ZR}(A)/2,$ and we also consider\n $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)$ as the sum of $\\mathcal{SQ}(A) + \\mathcal{ZR}(A)/2$ and $\\mathcal{ZR}(A)/2.$ Therefore, we begin by doing delicate estimates on both  $\\mathcal{SQ}(A) + \\mathcal{ZR}(A)/2$ and $\\mathcal{ZR}(A)/2.$ \n \n \\subsection{$\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2$ value}\n Let $A\\subset \\mathbb F_q^d.$\n Recall that $\\mathcal{ZR}(A)$ and $\\mathcal{SQ}(A)$ denote the numbers of pairs $(x, y)\\in A\\times A$ such that $||x- y||$ is zero and is a square number in $\\mathbb F_q^*$, respectively; namely $||x- y||=0$ and $||x-y||=r^2$ for some $r\\in \\mathbb F_q^*.$ Since $ r^2=(-r)^2$ for $r\\in \\mathbb F_q^*,$  we can write\n $$ \\mathcal{SQ}(A)= \\frac{1}{2} \\sum_{r\\in \\mathbb F_q^*} \\sum_{x, y\\in A: ||x-y||=r^2} 1.$$\n Since $\\mathcal{ZR}(A)=\\sum\\limits_{x,y\\in A: ||x-y||=0} 1$,  we have\n $$ \\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2} =\\frac{1}{2} \\sum_{r\\in \\mathbb F_q} \\sum_{x, y\\in A: ||x-y||=r^2} 1.$$\n \n \n Now notice that for each fixed $r\\in \\mathbb F_q,$ there are exactly $q$ pairs $(s,s')\\in \\mathbb F_q\\times \\mathbb F_q$ such that \n $r=s-s'.$\n This implies that\n \\begin{align*} \\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}&=  \\frac{1}{2q} \\sum_{s,s'\\in \\mathbb F_q}\n \\sum_{x, y\\in A: ||x-y||=(s-s')^2} 1\\\\\n &= \\frac{1}{2q} \\sum_{\\substack{(x,s), (y, s')\\in A\\times \\mathbb F_q:\\\\\n ||(x,s)- (y, s')||_{C_{d+1}}=0}} 1. \\end{align*}\n Put $E=A\\times \\mathbb F_q \\subset \\mathbb F_q^{d+1}$ and $ \\widetilde{x}=(x,s), \\widetilde{y}=(y, s')\\in \\mathbb F_q^{d+1}.$ It follows \n $$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}= \\frac{1}{2q} \\sum_{\\widetilde{x},\\widetilde{y}\\in E:\\widetilde{x}-\\widetilde{y}\\in C_{d+1}}1.$$\n We now use Lemma \\ref{AlexF} with $n=d+1$ to estimate the above summation. Then we see \n \\begin{equation}\\label{MainForm} \\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}= \\frac{1}{2q} q^{2(d+1)} \\sum_{\\widetilde{m}\\in \\mathbb F_q^{d+1}} \\widehat{C_{d+1}}(\\widetilde{m}) |\\widehat{E}(\\widetilde{m})|^2.\\end{equation}\nWe replace the above value $\\widehat{C_{d+1}}(\\widetilde{m})$ by  \\eqref{FFC} of Lemma \\ref{FTForm}. Then a direct computation gives us the following:\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|E|^2}{2q^2} +\\frac{q^{d-1} \\eta(-1) G_1^{d+1}}{2} \\sum_{\\widetilde{m}\\in \\mathbb F_q^{d+1}}  \\sum_{s\\ne 0} \\eta^{d+1}{(s)} \\chi\\left(\\frac{||\\widetilde{m}||_{C_{d+1}}}{-4s}\\right)  \n|\\widehat{E}(\\widetilde{m})|^2.$$\nNotice that $|E|=|A|q$ and $\\widehat{\\mathbb F_q}(t)=1$ for $t=0,$ and $0$ otherwise. For $\\widetilde{m}=(m,t)\\in \\mathbb F_q^d\\times \\mathbb F_q,$ we can check that\n$$\\widehat{E}(\\widetilde{m})=\\widehat{A\\times \\mathbb F_q} (m, t)=\\widehat{A}(m) \\widehat{\\mathbb F_q}(t).$$\nTherefore, we obtain\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1} \\eta(-1) G_1^{d+1}}{2} \\sum_{m\\in \\mathbb F_q^{d}}  \\sum_{s\\ne 0} \\eta^{d+1}{(s)} \\chi\\left(\\frac{||m||}{-4s}\\right)  \n|\\widehat{A}(m)|^2.$$\nBy a simple change of variables, $1/(-4s) \\to s,$ and the properties of $\\eta$,  \n\\begin{equation}\\label{A+F} \n\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1} \\eta^d(-1) G_1^{d+1}}{2} \\sum_{m\\in \\mathbb F_q^{d}}  \\sum_{s\\ne 0} \\eta^{d+1}{(s)} \\chi\\left({s||m||}\\right)|\\widehat{A}(m)|^2.  \n\\end{equation}\nFor better exposition, we will use the following notations.\n\\begin{definition} Let $A\\subset \\mathbb F_q^d.$ We define that\n$$\\Omega^+(A)=\\sum_{m\\in \\mathbb F_q^d: \\eta(||m||)=1} |\\widehat{A}(m)|^2,$$ \n$$\\Omega^-(A)=\\sum_{m\\in \\mathbb F_q^d: \\eta(||m||)=-1} |\\widehat{A}(m)|^2,$$ \nand\n$$\\Omega^0(A)=\\sum_{m\\in \\mathbb F_q^d: ||m||=0} |\\widehat{A}(m)|^2.$$\n\\end{definition}\nIt is clear that $\\Omega^+(A), \\Omega^-(A)$, and $\\Omega^0(A)$ are  non-negative real numbers. In addition, notice that if $A$ is a subset of $\\mathbb F_q^d,$ then\n\\begin{equation}\\label{triv} \\Omega^0 (A)\\ge |\\widehat{A}(0,\\ldots, 0)|^2 = q^{-2d} |A|^2.\\end{equation}\n\n Recall that the sphere $S_0$ in $\\mathbb F_q^d$ with zero radius, centered at the origin, is defined by \n$$ S_0:=\\{x\\in \\mathbb F_q^d: ||x||=0\\}.$$ With this notation and the definition of the Fourier transform, we have \n $$\\Omega^0(A)=\\sum_{m\\in S_0}|\\widehat{A}(m)|^2=q^{-d}\\sum_{x,y\\in A} \\widehat{S_0}(x-y).$$ \nBy the same argument as in the proof of Lemma \\ref{FTForm}, the Fourier transform of $S_0$, denoted by $\\widehat{S_0},$ is given as follows (see, for example, \\cite{IK10}): for $m\\in \\mathbb F_q^d,$ \n\\begin{equation}\\label{S0Fourier} \\widehat{S_0}(m)=\\frac{\\delta_0(m)}{q} + q^{-d-1} \\eta^d(-1) G_1^d \\sum_{s\\ne 0} \\eta^d(s) \\chi\\left(\\frac{||m||}{4s}\\right).\\end{equation}\nIn particular,  when $d\\ge 3$ is odd, one has\n$$ \\Omega^0(A)=q^{-d-1}|A| +q^{-2d-1}\\eta(-1)G_1^{d+1} \\sum_{x,y\\in A} \\eta(||x-y||).$$\nSince $\\Omega^0(A)$ is a real number and $|G_1|=q^{1/2},$ we see that if $d\\ge 3$ is odd, then\n$$\\Omega^0(A)\\le q^{-d-1}|A| +q^{-\\frac{3d+1}{2}} |A|^2.$$\nOne the other hand,  the Plancherel theorem gives the following estimate:\n$$ \\Omega^0(A) \\le \\sum_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 =q^{-d}|A|.$$\nIn conclusion, for any odd $d\\ge 3$, we get\n\\begin{equation}\\label{Omezero}\n\\Omega^0(A) \\le \\min\\{q^{-d}|A|, ~q^{-d-1}|A| +q^{-\\frac{3d+1}{2}} |A|^2\\}.\n\\end{equation}\nIn odd dimensions $d\\ge 3,$ we obtain the following formula.\n\\begin{proposition}\\label{pro4.2} Let $A$ be a set in $\\mathbb F_q^d.$\n\\begin{enumerate}\n\\item  If  $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ \nthen\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} -\\frac{q^{\\frac{3d+1}{2}}}{2}\\Omega^0(A)+ \\frac{q^{\\frac{d-1}{2}}|A|}{2}.$$\n\\item If $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{\\frac{3d+1}{2}}}{2}\\Omega^0(A)- \\frac{q^{\\frac{d-1}{2}}|A|}{2}.$$\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof} Since $d\\ge 3$ is odd,  the inequality  \\eqref{A+F} becomes\n$$ \\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1} \\eta(-1) G_1^{d+1}}{2} \\sum_{m\\in \\mathbb F_q^{d}}  \\sum_{s\\ne 0} \\chi\\left({s||m||}\\right)|\\widehat{A}(m)|^2. $$ \nBy the orthogonality of $\\chi$, we compute the sum over $s\\ne 0$. Then we get\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} \n+\\frac{q^{d} \\eta(-1) G_1^{d+1}}{2} \\sum_{m\\in  S_0} |\\widehat{A}(m)|^2 -\\frac{q^{d-1} \\eta(-1) G_1^{d+1}}{2} \\sum_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2.$$\nSince $\\sum\\limits_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2=q^{-d}|A|$, we have \n\\begin{equation}\n\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} \n+\\frac{q^{d} \\eta(-1) G_1^{d+1}}{2} \\Omega^0(A) -\\frac{\\eta(-1) G_1^{d+1}|A|}{2q}. \n\\end{equation}\nTo prove the first part (1) of the proposition, it suffices to show that $\\eta(-1)G_1^{d+1}= -q^{\\frac{d+1}{2}}.$ Since $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ the equation follows from Corollary \\ref{Corm} (1) with $n=d+1.$ \nTo prove the second part (2) of the proposition, it is enough to show that $\\eta(-1)G_1^{d+1}= q^{\\frac{d+1}{2}}.$ Notice that this equation  follows immediately from Corolalry \\ref{Corm} (2) with $n=d+1.$\n\\end{proof}\n\nIn even dimensions $d\\ge 2,$ we have\n\\begin{proposition} \\label{pro4.3}\nLet $A$ be a set in $\\mathbb F_q^d.$\n\\begin{enumerate}\n\\item If $d\\ge 2$ is even, then \n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1}  G_1^{d+2}}{2} \\Omega^+(A)-\\frac{q^{d-1}  G_1^{d+2}}{2} \\Omega^-(A).$$\n\\item In particular, if  $d\\equiv 2 \\mod{4},$ then\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^+(A)-\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^-(A). $$\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof}\nSince $d\\ge 2$ even,  \\eqref{A+F} implies that\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1}  G_1^{d+1}}{2} \\sum_{m\\in \\mathbb F_q^{d}}  \\sum_{s\\ne 0} \\eta{(s)} \\chi\\left({s||m||}\\right)|\\widehat{A}(m)|^2. $$ \nNote that the sum over $s\\ne 0$ is $G_1 \\eta(||m||).$ Since $\\eta(||m||)=0$ for $||m||=0,$ the first statement (1) of the proposition follows. Proposition \\ref{pro4.3} (2)  is a  direct consequence of Proposition \\ref{pro4.3} (1) since  $G_1^{d+2}=q^{(d+2)/2}$ for $d\\equiv 2 \\mod{4}.$\n\\end{proof}\n\n\n\\subsection{$\\mathcal{ZR}(A)/2$ value} \nRecall that for $A\\subset \\mathbb F_q^d,$ we have\n$$\\mathcal{ZR}(A)=\\sum_{x,y\\in A: ||x-y||=0} 1.$$ \nBy applying Lemma \\ref{AlexF} with  $n=d$ and $V=S_0,$ we have\n$$ \\frac{\\mathcal{ZR}(A)}{2}= \\frac{q^{2d}}{2} \\sum_{m\\in \\mathbb F_q^d} \\widehat{S_0}(m) |\\widehat{A}(m)|^2.$$\nThe following equality  can be easily obtained by a direct computation after replacing the above value $\\widehat{S_0}(m)$ by the value in \\eqref{S0Fourier}.\n\\begin{equation}\\label{defA0}\\frac{\\mathcal{ZR}(A)}{2}= \\frac{|A|^2}{2q}+ \\frac{q^{d-1} \\eta^d(-1) G_1^d }{2} \\sum_{m\\in \\mathbb F_q^d} \n\\left(\\sum_{s\\ne 0} \\eta^d(s) \\chi\\left(s||m||\\right)  \\right) |\\widehat{A}(m)|^2,\\end{equation}\nwhere we also used a simple change of variables, $1/(4s)\\to s,$ and the properties of the quadratic character $\\eta$ of $\\mathbb F_q.$\\\\\n\nIn odd dimensions $d\\ge 3,$  the value $\\mathcal{ZR}(A)/2$ is written as follows.\n\n\\begin{proposition}\\label{pro4.4} Let $A$ be a set in $\\mathbb F_q^d.$\n\\begin{enumerate}\n\\item  If  $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ \nthen\n$$ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2q} -\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^+(A)+\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^-(A). $$\n\\item If $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n$$\\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^+(A)-\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^-(A).$$\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof} Since $d$ is odd,  $\\eta^d=\\eta$ and so \nthe equality \\eqref{defA0} becomes\n\\begin{align*} \\frac{\\mathcal{ZR}(A)}{2}&= \\frac{|A|^2}{2q}+ \\frac{q^{d-1} \\eta(-1) G_1^d }{2} \\sum_{m\\in \\mathbb F_q^d} \n\\left(\\sum_{s\\ne 0} \\eta(s) \\chi\\left(s||m||\\right)  \\right) |\\widehat{A}(m)|^2\\\\\n&=\\frac{|A|^2}{2q}+ \\frac{q^{d-1} \\eta(-1) G_1^{d+1} }{2} \\sum_{m\\in \\mathbb F_q^d} \n \\eta(||m||) |\\widehat{A}(m)|^2.\\end{align*}\nSince $\\eta{(||m||)}=0$ for $||m||=0,$  it follows by the definitions of $\\Omega^+(A)$ and $\\Omega^-(A)$ that \n$$ \\frac{\\mathcal{ZR}(A)}{2} = \\frac{|A|^2}{2q}+ \\frac{q^{d-1} \\eta(-1) G_1^{d+1} }{2} \\Omega^+(A) - \\frac{q^{d-1} \\eta(-1) G_1^{d+1} }{2} \\Omega^-(A).$$\nApplying  Corollary \\ref{Corm} (1) and (2) with $n=d+1$,  we complete the proof.\n\\end{proof}\n\n\nIn even dimensions $d\\ge 2,$ we have\n\\begin{proposition} \\label{pro4.5}\nLet $A$ be a set in $\\mathbb F_q^d.$\n\\begin{enumerate}\n\\item  If  $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ then\n$$ \\frac{\\mathcal{ZR}(A)}{2}= \\frac{|A|^2}{2q} -\\frac{q^{\\frac{3d}{2}}}{2} \n\\Omega^0(A)+ \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n\n\n\\item If $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ then\n$$\\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^0(A)- \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n\\end{enumerate}\n\\end{proposition}\n\\begin{proof}\nSince $d$ is even, $\\eta^d\\equiv 1.$ Hence, the equality \\eqref{defA0} and the orthogonality of $\\chi$ yield \n\\begin{align*}\\frac{\\mathcal{ZR}(A)}{2}&= \\frac{|A|^2}{2q}+ \\frac{q^{d-1}  G_1^d }{2} \\sum_{m\\in \\mathbb F_q^d} \n\\left(\\sum_{s\\ne 0}  \\chi\\left(s||m||\\right)  \\right) |\\widehat{A}(m)|^2\\\\\n&=\\frac{|A|^2}{2q}+ \\frac{q^{d}  G_1^d }{2}\\Omega^0(A) -\\frac{q^{d-1}  G_1^d }{2} \\sum_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2.\\end{align*} \nNotice by the Plancherel theorem that $\\sum\\limits_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2=q^{-d}|A|.$ Then the proposition follows from the Gauss sum values given in  Corollary \\ref{Corm} (3) and (4) with $n=d.$\n\\end{proof}\n\n\\section{Proofs of results on $\\mathcal{SQ}(A)+\\mathcal{ZR}(A)$ (Theorem \\ref{mainN0})}\nTo get a good upper bound of $\\mathcal{SQ}(A)+\\mathcal{ZR}(A)$,  we will use the previously established estimates of $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2$ and $\\mathcal{ZR}(A).$ In other words,  we will find an upper bound of the sum of $(\\mathcal{SQ}(A)+\\mathcal{ZR}(A)/2)$ and $\\mathcal{ZR}(A)/2.$\n \n \\subsection{ Proof of Theorem \\ref{mainN0} (1)}\n Since we assume that $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ Propositions \\ref{pro4.2} (1) and \\ref{pro4.4} (1) play a key role in proving this theorem. \\\\\nBy inequality \\eqref{triv} and Proposition \\ref{pro4.2} (1), we see that\n\\begin{equation}\\label{TK1}\n(\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2)\\le \\frac{|A|^2}{2} -\\frac{|A|^2}{2q^{\\frac{d-1}{2}}}+ \\frac{q^{\\frac{d-1}{2}}|A|}{2}.\n\\end{equation}  \nIgnoring a negative term, Proposition \\ref{pro4.4} (1) implies that \n$$\\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^-(A). $$\nOne can check that\n$$\\Omega^{-}(A)\\le \\sum_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 -\\Omega^0(A) \\le q^{-d}|A|-|\\widehat{A}(0,\\ldots, 0)|^2=q^{-d}|A|-q^{-2d}|A|^2.$$ \nThus, we obtain\n$$\\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2q} +\\frac{q^{\\frac{d-1}{2}}|A|}{2} - \\frac{|A|^2}{2q^{\\frac{d+1}{2}}}.$$\n\nSince $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A) =(\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2)+ {\\mathcal{ZR}(A)}/{2},$ we complete the proof by adding the above two inequalities. \\qed\\\\\n\n\\subsection{Proof of Theorem \\ref{mainN0} (2)}\nSince $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ we are able to use the following result of Proposition \\ref{pro4.2} (2):\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{\\frac{3d+1}{2}}}{2}\\Omega^0(A)- \\frac{q^{\\frac{d-1}{2}}|A|}{2}.$$\nSince $\\Omega^0(A)$ is dominated by $\\sum\\limits_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 - \\Omega^+(A)=q^{-d}|A|-\\Omega^+(A),$  it follows that\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2} +\\frac{q^{\\frac{3d+1}{2}}}{2}(q^{-d}|A|-\\Omega^+(A))- \\frac{q^{\\frac{d-1}{2}}|A|}{2}. $$\n\nProposition \\ref{pro4.4} (2) implies that\n$$\\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^+(A).$$\nAdding the above two estimates, we have\n $$\\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q} + \\frac{q^{\\frac{d+1}{2}} |A|}{2} -\\frac{q^{\\frac{d-1}{2}} |A|}{2}-\\left( \\frac{q^{\\frac{3d+1}{2}}}{2}- \\frac{q^{\\frac{3d-1}{2}}}{2}\\right)\\Omega^+(A).$$\nSince the term containing $\\Omega^+(A)$ value is negative or zero, the theorem follows. \\qed\\\\\n\\subsection{Proof of Theorem \\ref{mainN0} (3)}\n Since $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$  the proof can use Propositions \\ref{pro4.3} (2) and \\ref{pro4.5} (1). \\\\\n Since $-\\Omega^-(A)\\le 0,$ Proposition \\ref{pro4.3} (2) implies that\n \\begin{equation}\\label{ThangDoowon3}\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2} +\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^+(A). \\end{equation} \n Combining this estimate with the fact that $\\Omega^+(A) \\le q^{-d}|A|-\\Omega^0(A)$, we obtain\n $$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2} +\\frac{q^{\\frac{3d}{2}}}{2} (q^{-d}|A|-\\Omega^0(A)).$$\nBy Proposition \\ref{pro4.5} (1), we have\n $$\\frac{\\mathcal{ZR}(A)}{2}= \\frac{|A|^2}{2q} -\\frac{q^{\\frac{3d}{2}}}{2} \n\\Omega^0(A)+ \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\nBy adding the above two inequalities, we have\n $$\\mathcal{SQ}(A)+\\mathcal{ZR}(A) \\le \\frac{|A|^2}{2} +\\frac{|A|^2}{2q}-q^{\\frac{3d}{2}} \n\\Omega^0(A)+ \\frac{q^{\\frac{d}{2}} |A|}{2} +\\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$ \nSince $-\\Omega^0(A)\\le -|\\widehat{A}(0,\\ldots,0)|^2=-q^{-2d}|A|^2,$  we obtain the statement of the theorem. \\qed\\\\\n\\subsection{Proof of Theorem \\ref{mainN0} (4)}\nSince $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ we may invoke Proposition \\ref{pro4.3} (1) and \\ref{pro4.5} (2). \\\\ \nFirst, Proposition \\ref{pro4.3} (1) tells us that for even integer $d\\ge 2,$\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2} +\\frac{q^{d-1}  G_1^{d+2}}{2} \\Omega^+(A)-\\frac{q^{d-1}  G_1^{d+2}}{2} \\Omega^-(A):=\\frac{|A|^2}{2}+ I + II.$$\nSince $d$ is  even,  Lemma \\ref{ExplicitGauss} implies that $G_1^{d+2}=\\pm q^{(d+2)/2}$ which is a real number. Moreover, both $ \\Omega^+(A)$ and $\\Omega^-(A)$ are non-negative real numbers. Hence, \none of  $I$ and $II$ is exactly non-negative and the other is non-positive. Therefore, in the case when  $G_1^{d+2}= q^{(d+2)/2},$ we have \n\\begin{equation}\\label{asseq1} \\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2}+ I=\\frac{|A|^2}{2}+\\frac{q^{\\frac{3d}{2}}}{2} \\Omega^+(A).\\end{equation}\nOn the other hand, when  $G_1^{d+2}= -q^{(d+2)/2},$ we have  \n\\begin{equation}\\label{asseq2}\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le\\frac{|A|^2}{2}+ II=\\frac{|A|^2}{2}+\\frac{q^{\\frac{3d}{2}}}{2} \\Omega^-(A).\\end{equation}\nIn any case, we have\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le\\frac{|A|^2}{2}+\\frac{q^{\\frac{3d}{2}}}{2} (q^{-d} |A|-\\Omega^0(A)),$$\nsince $\\max\\{\\Omega^+(A), ~\\Omega^-(A)\\} \\le  \\sum\\limits_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 -\\Omega^0(A) = q^{-d}|A|-\\Omega^0(A).$\\\\\nBy Proposition \\ref{pro4.5} (2), we have\n$$\\frac{\\mathcal{ZR}(A)}{2}=\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^0(A)- \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\nHence, adding the above two estimates, we obtain the required upper bound of $\\mathcal{SQ}(A)+ \\mathcal{ZR}(A).$ \\qed\\\\\n\\section{Proofs of results on $\\mathcal{SQ}(A)$ in odd dimensions (Theorem \\ref{mainNOdd})}\n\n\n\n\n\\subsection{Proof of Theorem \\ref{mainNOdd} (1)}\nSince the assumtion that $d\\equiv 3 \\mod{4}$ and $ q\\equiv 3 \\mod{4}$ is the same as that of Theorem\n\\ref{mainN0} (1), we will make use certain estimates given in the proof of Theorem \\ref{mainN0} (1).\\\\\nAs in \\eqref{TK1}, we have\n\\begin{equation}\\label{TK11}\n    (\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2)\\le \\frac{|A|^2}{2} -\\frac{|A|^2}{2q^{\\frac{d-1}{2}}}+ \\frac{q^{\\frac{d-1}{2}}|A|}{2}.\n\\end{equation}\nSince $\\Omega^-(A)$ is non-negative, Proposition \\ref{pro4.4} (1) implies that\n$$ -\\mathcal{ZR}(A)/2\\le  -\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2} \\Omega^+(A).  $$\n\nAs before, we can check that $\\Omega^+(A)$ is bounded by   $q^{-d}|A|-q^{-2d} |A|^2$ . Thus, we have\n\\begin{equation} \\label{TK2}\n-\\mathcal{ZR}(A)/2\\le  -\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2}(q^{-d}|A|-q^{-2d} |A|^2)= -\\frac{|A|^2}{2q}+\\frac{q^{\\frac{d-1}{2}}|A|}{2}-\\frac{|A|^2}{2q^{\\frac{d+1}{2}}}.\n\\end{equation}\n\nBy the definition of $\\mathcal{ZR}(A),$ it is clear that $\\mathcal{ZR}(A)\\ge |A|.$ Hence, we also have\n\\begin{equation}\\label{TK3} -\\mathcal{ZR}(A)/2\\le -|A|/2.\\end{equation}\n\nSince we can write $\\mathcal{SQ}(A)= (\\mathcal{SQ}(A)+ \\mathcal{ZR}(A)/2) -\\mathcal{ZR}(A)/2,$ the estimates \\eqref{TK11} and \\eqref{TK2} imply that\n$$\\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + q^{\\frac{d-1}{2}} |A| -\\frac{|A|^2}{2q}- \\frac{|A|^2}{2 q^{\\frac{d-1}{2}}} - \\frac{|A|^2}{2 q^{\\frac{d+1}{2}}}.$$\n\nCombining \\eqref{TK11} and \\eqref{TK3}, we also obtain that\n$$ \\mathcal{SQ}(A) \\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d-1}{2}} |A|}{2}\n  - \\frac{|A|^2}{2 q^{\\frac{d-1}{2}}} -\\frac{|A|}{2}.$$\n  \nComparing the above two estimates of $\\mathcal{SQ}(A),$ we obtain the statement of Theorem \\ref{mainNOdd} (1).\n\\qed \\\\\n\n\\subsection{Proof of Theorem \\ref{mainNOdd} (2)}\nSince  $d\\equiv 1 \\mod{4},$ or $d\\equiv 3 \\mod{4}$ and $q\\equiv 1 \\mod{4},$  we will make use of Propositions \\ref{pro4.2} (2) and \\ref{pro4.4} (2).\\\\\nCombining inequality \\eqref{Omezero} and Proposition \\ref{pro4.2} (2), we see that\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le\\frac{|A|^2}{2} +\\frac{q^{\\frac{3d+1}{2}}}{2}\\min\\{q^{-d}|A|, ~q^{-d-1}|A| +q^{-\\frac{3d+1}{2}} |A|^2\\}- \\frac{q^{\\frac{d-1}{2}}|A|}{2}.$$\n As seen in \\eqref{TK3},  we have\n $$  -\\mathcal{ZR}(A)/2\\le -|A|/2.$$\n Hence, the above two estimates imply that\n\\begin{equation}\\label{OE1} \n\\mathcal{SQ}(A) \\le \\frac{|A|^2}{2} - \\frac{q^{\\frac{d-1}{2}} |A|}{2}\n  -\\frac{|A|}{2}+ \\min\\left\\{ \\frac{q^{\\frac{d+1}{2}} |A|}{2},~ \\frac{q^{\\frac{d-1}{2}} |A|}{2}+ \\frac{|A|^2}{2} \\right\\}.\n\\end{equation} \nNext, we deduce another upper bound of $\\mathcal{SQ}(A).$ \nSince $\\Omega^0(A)=\\sum\\limits_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 -\\Omega^+(A)-\\Omega^-(A) =q^{-d}|A|-\\Omega^+(A)-\\Omega^-(A)\\le q^{-d}|A|-\\Omega^-(A),$ it follows by Proposition \\ref{pro4.2} (2) that\n\\begin{equation}\\label{Doowon2}\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2} +\\frac{q^{\\frac{3d+1}{2}}}{2}(q^{-d}|A|-\\Omega^-(A))- \\frac{q^{\\frac{d-1}{2}}|A|}{2}.\\end{equation}\nSince $-\\Omega^+(A)\\le 0,$  Proposition \\ref{pro4.4} (2) implies that \n$$-\\frac{\\mathcal{ZR}(A)}{2}\\le -\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d-1}{2}}}{2}\\Omega^-(A).$$\nSince $-\\Omega^-(A)\\le 0,$ adding the above two estimates gives us \n$$\\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d+1}{2}} |A|}{2} -\\frac{q^{\\frac{d-1}{2}} |A|}{2}-\\frac{|A|^2}{2q}.$$\nCombining this estimate with inequality \\eqref{OE1}, we complete the proof of Theorem \\ref{mainNOdd} (2).   \n\\qed\n\\\\\n\\section{Proofs of results on $\\mathcal{SQ}(A)$ in even dimensions (Theorem \\ref{mainNEven})}\nWe begin by proving an upper bound of $\\mathcal{SQ}(A)$ for even $d\\ge 2.$ We will use some certain estimates appearing in the proof of Theorem \\ref{mainN0}(4).\n\\begin{lemma}\\label{evenA+}\nLet $A$ be a subset of $\\mathbb F_q^d.$ If $d\\ge 2 $ is even, then we have\n$$ \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}}  }{2}|A|- \\frac{|A|^2}{2q^{\\frac{d}{2}}  }-\\frac{ |A| }{2}.$$\n\\end{lemma}\n\\begin{proof}\nAs observed in the proof of Theorem \\ref{mainN0} (4), either the inequality \\eqref{asseq1} or the inequality \\eqref{asseq2} happens. Without loss of generality, we can assume that  the inequality \\eqref{asseq1} holds. By the same method, we can deal with the other case.  As seen before, we have\n$\\Omega^+(A)\\le \\sum_{m\\in \\mathbb F_q^d} |\\widehat{A}(m)|^2 -\\Omega^0(A)\\le q^{-d}|A|-q^{-2d}|A|^2$.\nThus, we obtain by \\eqref{asseq1} that for even $d\\ge 2,$\n\\begin{equation}\\label{evenEasy}\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2} \\le \\frac{|A|^2}{2}+\\frac{q^{\\frac{3d}{2}}}{2} ( q^{-d}|A|-q^{-2d}|A|^2 )=\\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}}  }{2}|A|- \\frac{|A|^2}{2q^{\\frac{d}{2}}}.\\end{equation}\nSince $\\mathcal{ZR}(A)\\ge|A|$ by the definition of $\\mathcal{ZR}(A),$ we have\n$$ -\\frac{\\mathcal{ZR}(A)}{2} \\le -\\frac{|A|}{2}.$$\n The lemma follows by adding the above two estimates. \n\\end{proof}\n\n\\subsection{Proof of Theorem \\ref{mainNEven} (1)}\nSince  $d\\equiv 2 \\mod{4}$ and $ q\\equiv 3 \\mod{4},$ we can also use certain estimates which were already established in the proof of Theorem \\ref{mainN0} (3). As seen in \\eqref{ThangDoowon3}, we have\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2} +\\frac{q^{\\frac{3d}{2}}}{2}\\Omega^+(A).$$\n Using the fact that $\\Omega^+(A)\\le q^{-d}|A|-\\Omega^0{A},$  we obtain\n$$\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2}\\le \\frac{|A|^2}{2}+ \\frac{q^{\\frac{3d}{2}}}{2}(q^{-d}|A|-\\Omega^0{A})$$\nBy Proposition \\ref{pro4.5} (1), it follows that \n $$ -\\frac{\\mathcal{ZR}(A)}{2}= -\\frac{|A|^2}{2q} +\\frac{q^{\\frac{3d}{2}}}{2} \n\\Omega^0(A)- \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\n\nAdding the above two estimates gives the statement of the theorem. \\qed\\\\\n\n\\subsection{ Proof of Theorem \\ref{mainNEven} (2)}\nSince $d\\equiv 0 \\mod{4},$ or $d\\equiv 2 \\mod{4}$ and $q\\equiv 1 \\mod{4},$ we will use the estimate \\eqref{evenEasy} and Proposition \\ref{pro4.5} (2).\\\\\nSince $d$ is even, it follows by \\eqref{evenEasy} that\n \\begin{equation}\\label{ThangKoh4}\\mathcal{SQ}(A)+ \\frac{\\mathcal{ZR}(A)}{2} \\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}}  }{2}|A|- \\frac{|A|^2}{2q^{\\frac{d}{2}}}.\\end{equation}\n\nSince $-\\Omega^0(A) \\le -|\\widehat{A}(0,\\ldots, 0)|^2 = -q^{-2d}|A|^2,$ Proposition \\ref{pro4.5} (2) implies that\n$$-\\frac{\\mathcal{ZR}(A)}{2}\\le -\\frac{|A|^2}{2q} -\\frac{|A|^2}{2q^{\\frac{d}{2}}}+ \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$\nThus, adding the above two estimates gives us an upper bound of $\\mathcal{SQ}(A)$ as follows: \n$$  \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}} |A|}{2}-\\frac{|A|^2}{q^{\\frac{d}{2}}} -\\frac{|A|^2}{2q}+ \\frac{q^{\\frac{d-2}{2}}|A|}{2}.$$ \n\nLemma \\ref{evenA+} also gives the following upper bound of $\\mathcal{SQ}(A)$:\n$$ \\mathcal{SQ}(A)\\le \\frac{|A|^2}{2} + \\frac{q^{\\frac{d}{2}}  }{2}|A|- \\frac{|A|^2}{2q^{\\frac{d}{2}}  }-\\frac{ |A| }{2}.$$\nThe statement of the theorem follows by a direct comparison with two upper bounds of $\\mathcal{SQ}(A).$ \\qed\\\\\n\n\\section*{Acknowledgments}\nDoowon Koh was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MIST) (No. NRF-2018R1D1A1B07044469). T. Pham would like to thank to the VIASM for the hospitality and for the excellent working condition.\n\n\n\n\n\n", "meta": {"timestamp": "2022-05-03T02:30:21", "yymm": "2111", "arxiv_id": "2111.14076", "language": "en", "url": "https://arxiv.org/abs/2111.14076", "abstract": "Let $\\mathbb{F}_q$ be a finite field of order $q$. Iosevich and Rudnev (2005) proved that for any set $A\\subset \\mathbb{F}_q^d$, if $|A|\\gg q^{\\frac{d+1}{2}}$, then the distance set $\\Delta(A)$ contains a positive proportion of all distances. Although this result is sharp in odd dimensions, it is conjectured that the right exponent should be $\\frac{d}{2}$ in even dimensions. During the last 15 years, only some improvements have been made in two dimensions, and the conjecture is still wide open in higher dimensions. To fill the gap, we need to understand more about the structures of the distance sets, the main purpose of this paper is to provide some structural theorems on the distribution of square and non-square distances.", "subjects": "Number Theory (math.NT); Combinatorics (math.CO)", "title": "Structural theorems on the distance sets over finite fields", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9888419677654414, "lm_q2_score": 0.8519528076067262, "lm_q1q2_score": 0.8424466907171276}}
{"url": "https://arxiv.org/abs/2209.13199", "title": "Normal Bundles of Rational Normal Curves on Hypersurfaces", "abstract": "Let $C$ be the rational normal curve of degree $e$ in $\\mathbb{P}^n$, and let $X\\subset \\mathbb{P}^n$ be a degree $d\\ge 2$ hypersurface containing $C$. In previous work, I. Coskun and E. Riedl showed that the normal bundle $N_{C/X}$ is balanced for a general $X$. H. Larson studied the case of lines ($e=1$) and computed the dimension of the space of hypersurfaces for which $N_{C/X}$ has a given splitting type. In this paper, we work with any $e\\ge 2$. We compute explicit examples of hypersurfaces for all possible splitting types, and for $d\\ge 3$, we compute the dimension of the space of hypersurfaces for which $N_{C/X}$ has a given splitting type. For $d=2$, we give a lower bound on the maximum rank of quadrics with fixed splitting type.", "text": "\\section{Introduction}\n\nRational curves play an important role in the study of birational and arithmetic geometry of projective varieties. The local structure of the space of rational curves on a variety is determined by its normal bundle. In this paper, we study the possible splitting types of the normal bundle of rational normal curves on a hypersurface in $\\P^n$ and obtain explicit examples of hypersurfaces for each splitting type. Unless otherwise specified, we work over an algebraically closed field $K$ of arbitrary characteristic $p$.\n\nLet $X$ be a degree $d$ hypersurface in $\\P^n$ containing a smooth rational curve $C$. By the Birkhoff-Grothendieck's theorem, the normal bundle of $C$ on $X$, $N_{C/X}$, splits as a direct sum $N_{C/X}\\cong \\bigoplus_{i=1}^{n-2}\\mathcal{O}_{\\P^1}(a_i)$. The collection of integers $a_i$ is called the \\textit{splitting type} of $N_{C/X}$. We say a splitting type is \\textit{balanced} when $|a_i-a_j|\\le 1$ for all $i$ and $j$. We are interested in studying the possible splitting types when $C$ is the rational normal curve of degree $e$ in $\\P^n$. Additionally, given a splitting type $E_{\\vec{a}}=\\bigoplus_{i=1}^{n-2}\\mathcal{O}(a_i)$, we can study the space of hypersurfaces $X$ containing $C$ and for which $N_{C/X}\\cong \\bigoplus_{i=1}^{n-2}\\mathcal{O}(a_i)$. Let $X$ be given by a degree $d$ polynomial $F$ as $X = V(F)$, and define\n\\[ \\Sigma = \\{ F \\ | \\ X \\text{ is a degree } d \\text{ hypersurface smooth along } C \\}\\subset H^0(\\mathcal{O}_{\\P^n}(d)) \\]\nand\n\\[ \\Sigma_{\\vec{a}} = \\{ F\\in \\Sigma \\ | \\ N_{C/X}\\cong E_{\\vec{a}} \\}\\subset \\Sigma . \\]\nThe codimension of the locus of vector bundles on $\\P^1$ with a specified splitting type $E_{\\vec{a}}$ in the versal deformation space is given by $h^1(\\P^1, \\End(E_{\\vec{a}}))$ (see \\cite{C08} Lemma 2.4 and \\cite{CR18}). We call it the \\textit{expected codimension} for $\\Sigma_{\\vec{a}}$ in $\\Sigma$ and observe that\n\\[ h^1(\\End(E_{\\vec{a}})) = h^1(E_{\\vec{a}}^*\\otimes E_{\\vec{a}}) = \\sum_{\\{i,j | a_i-a_j\\le -2\\}}(a_j-a_i-1). \\]\n\nNormal bundles of rational curves have been studied in many different works (see \\cite{Sa80}, \\cite{GS80}, \\cite{EV81}, \\cite{EV82}, \\cite{M86}, \\cite{Ran}, \\cite{AlzatiRe}, \\cite{AlzatiReTortora}, \\cite{CR18}, \\cite{Ran20}, \\cite{Ran22}). In \\cite{CR18}, I. Coskun and E. Riedl show that the locus of rational curves in $\\P^n$ with a given splitting type has arbitrarily many components, and the difference between the expected dimension and actual dimension of a component can grow arbitrarily large as the degree of the curve increases.\n\nFor rational curves on hypersurfaces, H. Larson (\\cite{L21}, proof of Theorem 1.1) studies the case of lines and shows that $\\Sigma_{\\vec{a}}$ is smooth of the expected codimension. For rational normal curves of degree $e\\ge 2$, I. Coskun and E. Riedl prove (\\cite{CR} Corollary 3.8) that the normal bundle on a degree $d\\ge 2$ general hypersurface is balanced. Here, we examine more closely the case of rational normal curves on hypersurfaces, and find explicit examples of hypersurfaces $X$ for each possible splitting type of $N_{C/X}$.\n\nIn the section 2, we describe the normal bundle as the kernel in the sequence\n\\[ 0\\longrightarrow N_{C/X}\\longrightarrow \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}\\overset{\\psi_F}{\\longrightarrow} \\mathcal{O}(de)\\longrightarrow 0, \\]\nso the splitting type of $N_{C/X}$ must have the form of a rank $(n-2)$ direct sum\n\\[ N_{C/X}\\cong \\left(\\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i)\\right )\\oplus \\left (\\bigoplus_{j=e+1}^{n}\\mathcal{O}(e-b_j) \\right ), \\]\nwith $a_i, b_j\\ge 0$, and degree $\\deg N_{C/X} = \\deg N_{C/\\P^n} - \\deg \\mathcal{O}(de) = e(n-d+1)-2$, equivalently $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^n b_j = e(d-1)-2$.  As a direct consequence of the surjectivity of the map $\\phi : H^0(\\mathcal{I}_C(d))\\to \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))$ for $d\\ge 3$ from \\cite{CR} (see section 2), we show that all these splitting types are achieved, and we look for explicit examples. We first examine the case $d\\ge 3$:\n\n\\begin{theorem}\\label{existencetheorem} (cf. Theorem \\ref{existence})\nLet $e\\le n$, $d\\ge 3$, and let $C$ be the rational normal curve of degree $e$ in $\\P^n$. Then for all splitting types \n\\[ E = \\left(\\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i)\\right )\\oplus \\left (\\bigoplus_{j=e+1}^{n}\\mathcal{O}(e-b_j) \\right ), \\]\nwith $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^n b_j = e(d-1)-2$, we obtain explicit examples of degree $d$ hypersurfaces $X$, smooth along the curve $C$, with normal bundle $N_{C/X}\\cong E$.\n\\end{theorem}\n\nWhen $d\\ge 4$, Bertini's Theorem implies that the general example is smooth.\n\\begin{corollary}\\label{existsmooth} (cf. Corollary \\ref{smoothexample})\n(char $K=0$) Let $d\\ge 4$, and assume the base field $K$ has characteristic $0$. For all splitting types $E$ as in the theorem, there exists a smooth hypersurface $X$ of degree $d$ containing the curve $C$ with normal bundle $N_{C/X}\\cong E$.\n\\end{corollary}\n\nIn addition, for $d\\ge 3$, we compute the dimension of $\\Sigma_{\\vec{a}}$ and show that, when $e=n$ or the splitting type does not have terms of degree $e+2$, we get the expected codimension in $\\Sigma$. When the dimension is not the expected one, we can compute the difference to its actual dimension. In particular, we show the difference can get arbitrarily large as $n$ grows.\n\n\\begin{theorem} (cf. Theorem \\ref{teoremasigmaa})\nLet $e\\le n$ and $d\\ge 3$. Given a splitting type $E_{\\vec{a}}$ as in Theorem \\ref{existencetheorem}, the locus $\\Sigma_{\\vec{a}}$ is irreducible and smooth of codimension $h^1(\\End(E_{\\vec{a}})) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n}))$ in $\\Sigma$.\n\\end{theorem}\n\n\\begin{corollary} (cf. Corollary \\ref{difference}) Let $z = |\\{ i \\ | \\ a_i = e+2 \\}|$ be the number of terms $\\mathcal{O}(e+2)$ in the splitting type $E_{\\vec{a}}$. Then the codimension of $\\Sigma_{\\vec{a}}$ in $\\Sigma$ is $h^1(\\End(E_{\\vec{a}})) - (n-e)z$.\n\\end{corollary}\n\nThe case $d=2$ presents additional difficulties, as for example, the map $\\phi : H^0(\\mathcal{I}_C(2))\\to \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(2e))$ is not surjective, so it is not immediately clear that all splitting types are achievable. Nonetheless, we can adapt the computation from the case $d\\ge 3$, and find explicity examples of quadrics for every possible splitting type.\n\n\\begin{theorem}\\label{quadricsexamples} (cf. Theorem \\ref{quadricprop1})\nFor any given splitting type \n\\[ E = \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right ), \\] with $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^nb_j = e-2$, we produce a quadric $X=V(F)$, smooth along $C$, for which $N_{C/X}\\cong E$.\n\nMore explicitly, for $e=n$, rearrange the $a_i$ in increasing order $a_1\\le \\cdots \\le a_{n-2}$, and let $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $i=1, \\cdots , n-2$. Then the quadric $X$ given by the polynomial\n\\[ F = \\sum_{i=0}^{n-2}Q_{\\beta_i+1,\\beta_i+2} =  Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + Q_{\\beta_2 + 1, \\beta_2 + 2} \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}, \\]\nwhere $Q_{ij} = x_ix_{j-1}-x_{i-1}x_j$, has normal bundle $N_{C/X}\\cong E$.\n\nFor $e<n$, let $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $1\\le i\\le e-2$. Let also $\\gamma_n = e$ and $\\gamma_{j} = e - b_n - b_{n-1} - \\cdots - b_{j+1}$ for $e+1\\le j\\le n-1$. Then a quadric $X$ such that $N_{C/X}\\cong E$ is given by\n\\begin{align*}\n    F = & \\sum_{i=0}^{e-2}Q_{\\beta_i+1,\\beta_i+2} + \\sum_{j=e+1}^n x_{\\gamma_j}x_j\\\\\n    = & (Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}) + (x_{\\gamma_j}x_{e+1} + \\cdots + x_{\\gamma_{n-1}}x_{n-1} + x_ex_n).\n\\end{align*}\n\n\\end{theorem}\n\n\n\n\nThe quadrics constructed in \\ref{quadricsexamples} are smooth along the curve $C$ but not necessarily smooth quadrics. We can, however, work on their quadratic form matrices to produce examples with a smaller singular locus. This allows us to find a lower bound for the maximum rank of quadrics with a given normal bundle:\n\n\\begin{theorem}\\label{quadricstheorem} (cf. Theorem \\ref{corankexample})\nFor every splitting type \n\\[ E = \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right ), \\]\nwith $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^nb_j = e-2$, we obtain an example of quadric $X$ of corank at most $\\sum _{a_i\\ge 4}(a_i-3)$ with $N_{C/X}\\cong E$. In particular, if $a_i\\le 3$ for all $i$, we find a smooth quadric $X$ with $N_{C/X}\\cong E$.\n\\end{theorem}\n\n\n\\noindent \\textbf{Organization of the paper.} In Section 2, we discuss definitions and computations that will be used throughout the paper. In Section 3, we work on the case $d\\ge 3$, both providing the example of a hypersurface with $N_{C/X}\\cong E_{\\vec{a}}$ and computing the dimension of $\\Sigma_{\\vec{a}}$. Section 4 is dedicated to the case $d=2$; we compute the example with a given splitting type, and prove Theorem \\ref{quadricstheorem}.\n\n\\medskip\n\n\\noindent \\textbf{Acknowledgements.} I would like to thank my advisor Izzet Coskun for our many discussions and his guidance throughout the writing of this paper. I would also like to thank Benjamin Gould, Yeqin Liu, Eric Riedl and Geoffrey Smith for reading my drafts and giving valuable suggestions.\n\n\n\n\\section{Preliminaries}\n\n  In this section we cover the definitions and results necessary to describe our approach. Most results here are proved in \\cite{CR}. We describe the rational normal curves, and give an explicit computation of their normal bundle on a given hypersurface.\n  \n\\medskip\n\n\\subsection{Rational Normal Curves} For $e\\le n$, we say the rational normal curve $R_e$ of degree $e$ in $\\P^n$ is the curve defined by\n\\[ (s^e : s^{e-1}t : s^{e-2}t^2 : \\cdots : st^{e-1} : t^e : 0 : \\cdots : 0): \\P^1\\to \\P^n. \\]\n\nIts homogeneous ideal $I_{R_e}\\subset k[x_0, \\cdots , x_n]$ is cut out by quadrics and linear forms:\n\\[ I_{R_e} = ( \\{ Q_{ij} = x_ix_{j-1} - x_{i-1}x_j \\mid 1\\le i < j\\le e \\}\\cup \\{x_{e+1}, \\cdots , x_n \\} ). \\]\n\nThe $Q_{ij}$ correspond to the $2\\times 2$ minors of the matrix\n\\[ \\begin{pmatrix}\nx_1 & x_2 & \\cdots & x_e\\\\\nx_0 & x_1 & \\cdots & x_{e-1}\n\\end{pmatrix}. \\]\n\nWe can get the relations between the $Q_{ij}$ as the $3\\times 3$ minors of the matrices\n\\[ \\begin{pmatrix}\nx_1 & x_2 & \\cdots & x_e\\\\\nx_1 & x_2 & \\cdots & x_e\\\\\nx_0 & x_1 & \\cdots & x_{e-1}\n\\end{pmatrix} \\text{ and } \\begin{pmatrix}\nx_0 & x_1 & \\cdots & x_{e-1}\\\\\nx_1 & x_2 & \\cdots & x_e\\\\\nx_0 & x_1 & \\cdots & x_{e-1}\n\\end{pmatrix}. \\]\n\nLet $R_n$ be the rational normal curve of degree $n$ in $\\P^n$. From the relations above, Proposition 2.4 in \\cite{CR} shows that the quadrics $Q_{i,i+1}$, for $1\\le i\\le n-1$ suffice to determine the elements of $H^0(N_{R_n/\\P^n})$.\n\n\\begin{proposition}\\label{relations}(\\cite{CR} Proposition 2.4)\nAn element $\\alpha \\in H^0(N_{R_n/\\P^n}) = \\Hom (\\mathcal{I}_{R_n/\\P^n}, \\mathcal{O}_C)$ is determined by the images $\\alpha (Q_{i,i+1})$, for $1\\le i\\le n-1$. Furthermore, $s^{n-i-1}t^{i-1}$ divides $\\alpha (Q_{i,i+1})$ and this is the only constraint on $\\alpha (Q_{i,i+1})$. If $b_{i,i+1}$, for $1\\le i\\le n-1$, are arbitrary polynomials of degree $n+2$, there exists an element $\\alpha \\in H^0(N_{R_n/\\P^n})$ such that $\\alpha (Q_{i,i+1}) = s^{n-i-1}t^{i-1}b_{i,i+1}$.\n\nIn addition, the image $\\alpha (Q_{i,j})$ of the other generators of $I_{R_n}$ are expressed in terms of $b_{i,i+1}$ by\n\\[ \\alpha (Q_{i,j}) = \\sum _{l=1}^{j-1} s^{n-j-i+l}t^{j+i-l-2}b_{l,l+1}. \\]\n\\end{proposition}\n\n\\begin{corollary}\\label{normalbundlee} (\\cite{CR} Corollary 2.6)\nFor an integer $e\\le n$, the normal bundle $N_{R_e/\\P^n}$ is $N_{R_e/\\P^e}\\oplus N_{\\P^e/\\P^n}\\cong \\mathcal{O}_{\\P^1}(e+2)^{e-1}\\oplus \\mathcal{O}_{\\P^1}(e)^{n-e}$.\n\\end{corollary}\n\n\\subsection{Normal bundles on hypersurfaces} Let $C$ be a smooth rational curve of degree $e$ and let $X$ be a degree $d$ hypersurface in $\\P^n$ containing $C$. Using the identification $N_{X/\\P^n}\\cong \\mathcal{O}_{X}(d)$, we can write the standard normal bundle sequence as\n\\[ 0\\longrightarrow N_{C/X}\\longrightarrow N_{C/\\P^n}\\overset{\\psi}{\\longrightarrow} N_{X/\\P^n}|_C\\cong \\mathcal{O}_{\\P^1}(de). \\]\n\nThus, we get a map\n\\[ \\phi : H^0(\\mathcal{I}_C(d))\\to \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de)) \\]\nthat sends polynomials $F$ defining $X = V(F)$ to elements $\\psi$ in $\\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))$.\n\nThis can be identified via the sequence\n\\[ 0\\longrightarrow \\mathcal{I}_{C/\\P^n}^2\\longrightarrow \\mathcal{I}_{C/\\P^n}\\longrightarrow N_{C/\\P^n}^*\\longrightarrow 0 \\]\nwhen we twist it by $\\mathcal{O}_{\\P^n}(d)$ and take global sections:\n\\begin{equation*}\\label{phisequence}\n0\\longrightarrow H^0(\\mathcal{I}_{C/\\P^n}^2(d))\\longrightarrow H^0(\\mathcal{I}_{C/\\P^n}(d))\\overset{\\phi}{\\longrightarrow} H^0(N_{C/\\P^n}^*(d)).\n\\end{equation*}\n\nWhen $C = R_e$ is the rational normal curve of degree $e$ in $\\P^n$ and $d\\ge 3$, I. Coskun and E. Riedl showed in \\cite{CR} that the map $\\phi$ is surjective, that is, every element $\\psi \\in \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))$ can be obtained as the map of a normal bundle sequence for some hypersurface $X$. It is not surjective when $d=2$. In the section 4, we show $\\phi$ is injective for $d=2$ and $e=n$, and we compute the dimension of its image.\n\nThe relations given in Proposition \\ref{relations} allow us to explicitly write the map $\\psi _F: N_{C/\\P^n}\\to \\mathcal{O}_{\\P^1}(de)$ for a hypersurface $X=V(F)$. First, let $e=n$. Write $F$ in terms of the generators of $I_C$, $F = \\sum _{i,j}F_{i,j}Q_{i,j}$. Then $\\psi_F(\\alpha) = \\sum_{i,j}F_{i,j}|_C\\alpha(Q_{i,j})$. By the relation from Proposition \\ref{relations}, we get\n\\[ \\psi_F(\\alpha) = \\sum _{i,j}F_{i,j}|_C\\sum _{l=i}^{j-1} s^{n-j-i+l}t^{j+i-l-2}b_{l,l+1}. \\]\n\nCollect the terms and write the sum as $\\sum_{i=1}^{n-1}c_ib_{i,i+1}$, then the map $\\psi_F: \\mathcal{O}_{\\P^1}(n+2)^{n-1}\\to \\mathcal{O}_{\\P^1}(dn)$ is given by the matrix $(c_1 \\ \\cdots \\ c_{n-1})$.\n\nWhen $e<n$, by Corollary \\ref{normalbundlee} the normal bundle $N_{C/\\P^n}$ splits as the direct sum $N_{C/\\P^n}\\oplus N_{\\P^e/\\P^n}$. We write $F = \\sum _{i,j}F_{i,j}Q_{i,j} + \\sum _{k=e+1}^nG_kx_k$, and collect the coefficients $c_1, \\cdots , c_{e-1}$ of the $b_{l,l+1}$ as above. Then the map $\\psi_F: \\mathcal{O}_{\\P^1}(e+2)^{e-1}\\oplus \\mathcal{O}_{\\P^1}(e)^{n-e}\\to \\mathcal{O}_{\\P^1}(de)$ is given by the matrix $(c_1 \\ \\cdots \\ c_{e-1}; \\ G_{e+1}|_C \\ \\cdots \\ G_n|C)$.\n\n\\medskip\n\nWe can use this description to obtain the map $\\psi_F$ from explicit hypersurfaces $X=V(F)$. By choosing the appropriate $\\psi_F$, we will then obtain each possible splitting type for $N_{C/X}$.\n\nFirst, observe that when $C$ is the rational normal curve of degree $e$ in $\\P^e$, the restrictions $F|_C$ of hypersurfaces of degree $k\\ge 1$ cut out the complete linear series $|\\mathcal{O}_C(k)|$, that is, rational normal curves are projectively normal.\n\n\\begin{lemma}\\label{projectivelynormal}\\cite{ACGH}\nFor every $k\\ge 1$, the map $H^0(\\mathcal{O}_{\\P^n}(k))\\to H^0(\\mathcal{O}_{C}(k))\\cong H^0(\\mathcal{O}_{\\P^1}(ek))$, $F\\mapsto F|_C$, is surjective.\n\\end{lemma}\n\nTo obtain an explicit $F$ for each polynomial in $H^0(\\mathcal{O}_{\\P^1}(ek))$ it suffices to write each monomial as a product of $k$ monomials of degree $e$. For instance, for $F|_C = s^{ek-2}t^2$ we can write $s^{ek-2}t^2 = s^{e(k-1)}(s^{e-2}t^2)$ and choose $F = x_0^{k-1}x_2$.\n\nBefore we advance to the general case, let us compute an example of a hypersurface $X$ for which $N_{C/X}$ is balanced in order to display the computation process with a simpler notation. In particular, we recover Corollary 3.8 of \\cite{CR} for the rational normal curve of degree $n$ in $\\P^n$. We make use of the same method in the proof of Theorem \\ref{existence}.\n\n\\begin{proposition}\\label{balancedexample} (\\cite{CR}, Corollary 3.8)\nLet $d\\ge 2$, and let $X$ be a general hypersurface of degree $d$ containing the rational normal curve $C$ of degree $n$ in $\\P^n$. Then the normal bundle $N_{C/X}$ is balanced.\n\\end{proposition}\n\\begin{proof}\nSince being balanced is an open condition in a family of vector bundles on $\\P^1$ with fixed rank and degree, it suffices to show an example of hypersurface $X$ for each $d$. The idea is to compute the kernel of $\\psi_F: \\mathcal{O}(n+2)^{n-1}\\to \\mathcal{O}(dn)$ directly from the linear relations between the entries of $\\psi_F$, which we call them the \\textit{column relations} of $\\psi_F$.\n\nIf $d=2$, consider $F = \\sum_{i=1}^{n-1} Q_{i,i+1}$, then\n\\[ \\psi_F = ( s^{n-2},\\ s^{n-3}t, \\cdots ,\\ st^{n-3},\\ t^{n-2} ). \\]\n\nThe columns $C_1, \\cdots , C_{n-1}$ of $\\psi_F$, satisfy the relations\n\\[ tC_i - sC_{i+1} = t(s^{n-1-i}t^{i-1}) - s(s^{n-2-i}t^{i}) = 0, \\ \\ 1\\le i\\le n-2. \\]\n\nThese relations define the vectors $K_i = (a_1, \\cdots , a_{n-1})$ with $a_i = t$, $a_{i+1} = -s$ and $a_j = 0$ for $j\\neq i, i+1$. Let $K$ be the matrix whose columns are $K_i$:\n\\[ K = \\begin{pmatrix}\nt & 0 & 0 & \\cdots & 0 & 0\\\\\n-s & t & 0 & \\cdots & 0 & 0\\\\\n0 & -s & t & \\cdots & 0 & 0\\\\\n\\vdots &  \\vdots & \\vdots & \\ddots & \\vdots & \\vdots\\\\\n0 &  0  & 0 & \\cdots & t & 0\\\\\n0 &  0  & 0 & \\cdots & -s & t\\\\\n0 &  0  & 0 & \\cdots & 0 & -s\\\\\n\\end{pmatrix}\\]\n\nThen $K$ defines a map\n\\[ K: \\mathcal{O}(n+1)^{n-2}\\to \\mathcal{O}(n+2)^{n-1} \\]\nwhose image, by the column relations above, is in the kernel $N_{C/X}$ of $\\psi_F$. Notice also that $K$ has maximum rank, so the map is injective, and thus it factors through the $N_{C/X}$. As $N_{C/X}$ has rank $n-2$, we have $N_{C/X}\\cong \\mathcal{O}(n+1)^{n-2}$.\n\nNow, let $d\\ge 3$. First, observe that if $N_{C/X}$ is balanced, then it must be  $\\mathcal{O}(A)^{m}\\oplus \\mathcal{O}(A+1)^{n-2-m}$ with\n\\[ A = \\left \\lfloor \\dfrac{n(n-d+1)-2}{n-2} \\right \\rfloor \\text{ and } m = A(n-2)-n(n-d). \\]\n\nThe approach is to construct $\\psi_F$ such that its column relations define an injective map\n\\[ K: \\mathcal{O}(A)^{m}\\oplus \\mathcal{O}(A+1)^{n-2-m}\\to \\mathcal{O}(n+2)^{n-1} \\]\nwhich, as in the case $d=2$, will be the kernel of $\\psi_F$. This means we want to obtain $m$ relations of degree $(n+2)-A$ and $n-2-m$ relations of degree $(n+2)-(A+1)$ between the columns of $\\psi_F$. To simplify the analysis, we look at polynomials $F = \\sum_{i=1}^{n-1}F_iQ_{i,i+1}$ with $F_i$ monomials. Set $F_1 = x_0^{d-2}$ and $F_{n-1} = x_n^{d-2}$, so $F$ induces a map of the form\n\\[ \\psi_F = ( (s^{n-2})(s^{n(d-2)}), \\ (s^{n-3}t)F_2|_C, \\cdots ,\\ (st^{n-3})F_{n-2}|_C, \\ (t^{n-2})(t^{n(d-2)}) ). \\]\n\nSince an arbitrary degree $d-2$ homogeneous polynomial $F_i$ has the form $F_i|_C = s^{n(d-2)-\\beta_i}t^{\\beta_i}$ when restricted to $C$, the map $\\psi_F$ has the form\n\\[ \\psi_F = ( (s^{n-2})(s^{n(d-2)}), \\ (s^{n-3}t)(s^{n(d-2)-\\beta_2}t^{\\beta_2}), \\cdots ,\\ (st^{n-3})(s^{n(d-2)-\\beta_{n-2}}t^{\\beta_{n-2}}), \\ (t^{n-2})(t^{n(d-2)}) ). \\]\n\nTo further simplify, we look for relations that only involve two consecutive entries of $\\psi_F$. One way to find such $\\psi_F$ is to choose the $\\beta_i$ in increasing order $0\\le \\beta_2\\le \\cdots \\le \\beta_{n-2}\\le n(d-2)$. This way, we get the columns relations\n\\begin{align*}\n    & t^{\\beta_2+1}C_1 - s^{\\beta_2+1}C_2 = 0, \\\\\n    & t^{\\beta_{i+1}-\\beta_i+1}C_i - s^{\\beta_{i+1}-\\beta_i+1}C_{i+1} = 0, \\ \\ 2\\le i\\le n-3 \\\\\n    & t^{n(d-2)-\\beta_{n-2}+1}C_{n-2} - s^{n(d-2)-\\beta_{n-2}+1}C_{n-1} = 0.\n\\end{align*}\n\nAnd as before, we define the matrix $K$ whose columns follow from the relations above:\n\\[ K = \\begin{pmatrix}\nt^{\\beta_2+1} & 0 & 0 & \\cdots & 0 & 0\\\\\n-s^{\\beta_2+1} & t^{\\beta_3-\\beta_2+1} & 0 & \\cdots & 0 & 0\\\\\n0 & -s^{\\beta_3-\\beta_2+1} & t^{\\beta_4-\\beta_3+1} & \\cdots & 0 & 0\\\\\n\\vdots &  \\vdots & \\vdots & \\ddots & \\vdots & \\vdots\\\\\n0 &  0  & 0 & \\cdots & t^{\\beta_{n-2}-\\beta_{n-3}+1} & 0\\\\\n0 &  0  & 0 & \\cdots & -s^{\\beta_{n-2}-\\beta_{n-3}+1} & t^{n(d-2)-\\beta_{n-2}+1}\\\\\n0 &  0  & 0 & \\cdots & 0 & -s^{n(d-2)-\\beta_{n-2}+1}\\\n\\end{pmatrix}\\]\n\nSince $K$ has maximum rank, it defines an injective map \n\\[ \\mathcal{O}(n+2-(\\beta_2+1))\\oplus \\left (\\bigoplus_{i=2}^{n-3}\\mathcal{O}(n+2-(\\beta_{i+1}+\\beta_i+1))\\right )\\oplus \\mathcal{O}(n+2-(n(d-2)-\\beta_{n-2}+1))\\overset{K}{\\to}\\mathcal{O}(n+2)^{n-1} \\]\nwhich, due to the relations, factors through the kernel of $\\psi_F$. Since the direct sum above and $N_{C/X}$ have the same rank, it follows that\n\\[ N_{C/X}\\cong \\mathcal{O}(n+2-(\\beta_2+1))\\oplus \\left (\\bigoplus_{i=2}^{n-3}\\mathcal{O}(n+2-(\\beta_{i+1}+\\beta_i+1))\\right )\\oplus \\mathcal{O}(n+2-(n(d-2)-\\beta_{n-2}+1)). \\]\n\nTo conclude, we pick the appropriate values for $\\beta_i$, $2\\le i\\le n-2$. Choose\n\\[ \\beta _i = \\left\\{\\begin{matrix}\n(i-1)(n+1-A), & 2\\le i\\le m+1\\\\ \n(i-1)(n-A)+m, & m+2\\le i\\le n-2\n\\end{matrix}\\right., \\]\nnoting that, by Lemma \\ref{projectivelynormal}, there exist $F_i$ such that $F_i|_C = s^{n(d-2)-\\beta_i}t^{\\beta_i}$. A simple computation then shows that the direct sum is $\\mathcal{O}(A)^{m}\\oplus \\mathcal{O}(A+1)^{n-2-m}$.\n\n\\end{proof}\n\n\\section{Hypersurfaces of Degree at Least 3}\n\nThroughout this section, we assume $d\\ge 3$. We study the splitting types of $N_{C/X}$ for hypersurfaces of degree $d$ in $\\P^n$ containing a degree $e\\le n$ rational normal curve $C = R_e$. By Corollary \\ref{normalbundlee}, the normal bundle $N_{C/X}$ is the kernel in the sequence\n\\[ 0\\longrightarrow N_{C/X}\\longrightarrow \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}\\overset{\\psi_F}{\\longrightarrow} \\mathcal{O}(de)\\longrightarrow 0 \\]\nfor $F$ a degree $d$ polynomial defining $X = V(F)$. \n\nThe surjectivity of the map $\\phi : H^0(\\mathcal{I}_{C/\\P^n}(d))\\to \\Hom (N_{C/\\P^n}, \\mathcal{O}(de))$ for $d\\ge 3$ (\\cite{CR} Theorem 3.1) implies the existence of hypersurfaces $X$ for any possible splitting type $E_{\\vec{a}}$: consider an injection $E_{\\vec{a}}\\to N_{C/\\P^n}$ and compute its cokernel $\\psi: N_{C/\\P^n}\\to \\mathcal{O}(de)$. Since $\\phi$ is surjective, there is a polynomial $F\\in H^0(\\mathcal{I}_{C/\\P^n}(d))$ that defines $\\psi$, and thus it induces $E_{\\vec{a}}$.\n\nHere, we look for examples of each possible splitting type for $N_{C/X}$, and the dimension of the space $\\Sigma_{\\vec{a}}$ of hypersurfaces $X$ with a given splitting type $E_{\\vec{a}}$.\n\nThe surjectivity of $\\phi$ is also enough to show that $N_{C/X}$ is balanced for the general $X$ (\\cite{CR}, Corollary 3.8), and it is a result we use to compute the dimension of $\\Sigma_{\\vec{a}}$.\n\n\\subsection{Examples of hypersurfaces for each splitting type}\n\nFrom the short exact sequence, a candidate for splitting type of $N_{C/X}$ must have the form of a rank $(n-2)$ direct sum \n\\[ N_{C/X}\\cong \\left(\\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i)\\right )\\oplus \\left (\\bigoplus_{j=e+1}^{n}\\mathcal{O}(e-b_j) \\right ), \\]\nwith $a_i, b_j\\ge 0$, and degree $\\deg N_{C/X} = \\deg N_{C/\\P^n} - \\deg \\mathcal{O}(de) = e(n-d+1)-2$, equivalently $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^n b_j = e(d-1)-2$. \n\n\\medskip\nWe start by finding examples of hypersurfaces $X$ smooth along $C$ for each splitting type. The proof of the theorem is a generalization of the computation from Proposition \\ref{balancedexample}.\n\n\\begin{theorem}\\label{existence} For all splitting types\n\\[ E = \\left(\\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i)\\right )\\oplus \\left (\\bigoplus_{j=e+1}^{n}\\mathcal{O}(e-b_j) \\right ), \\]\nwith $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^n b_j = e(d-1)-2$, we obtain explicit examples of degree $d$ hypersurfaces $X$, smooth along the curve $C$, with normal bundle $N_{C/X}\\cong E$.\n\\end{theorem}\n\n\\begin{proof} \nWe divide the proof into the cases $e=n$ and $e<n$.\n\nFor $e=n$, we ask if we can get all splitting types $E = \\bigoplus_{i=1}^{n-2}\\mathcal{O}(n+2-a_i)$ with $a_i\\ge 0$ and $\\sum _{i=1}^{n-2}a_i = n(d-1)-2$. The approach is to construct a map $\\psi_F$ from which we can easily compute the kernel from the linear relations between its entries, which we call the \\textit{column relations} of $\\psi_F$.\n\nIn order to start with a simpler $\\psi_F$, consider polynomials $F$ of the form $F = \\sum_{i=1}^{n-1}F_iQ_{i,i+1}$ with $F_1 = x_0^{d-2}$ and $F_{n-1} = x_n^{d-2}$, so the map $\\psi_F: \\mathcal{O}(n+2)^{n-1}\\to \\mathcal{O}(dn)$ has the form\n\\[ \\psi_F = (s^{n-2}(s^{n(d-2)}), \\ \\ (s^{n-3}t)F_2|_C, \\ \\cdots ,\\ (st^{n-3})F_{n-2}|_C, \\ \\ t^{n-2}(t^{n(d-2)})). \\]\n\nFirst, we check that these are all smooth along the curve $C$. Taking partial derivatives $\\frac{\\partial F}{\\partial x_i}$ and restricting to the curve (that is, taking $Q_{ij} = 0$), we obtain\n\\[ \\frac{\\partial F}{\\partial x_0} = -x_0^{d-2}x_2, \\ \\ \\frac{\\partial F}{\\partial x_{2}} = -x_0^{d-1} + 2x_{2}F_{2} - x_{4}F_{3} \\ \\ \\text{ and } \\ \\ \\frac{\\partial F}{\\partial x_{n-2}} = -x_n^{d-1} + 2x_{n-2}F_{n-2} - x_{n-4}F_{n-3}. \\]\nIf $F$ is singular at a point $P = (s^n : s^{n-1}t : \\cdots : t^n)$ on $C$, then $\\frac{\\partial F}{\\partial x_0} = 0$ gives $s=0$ or $t=0$. If $s=0$, then $\\frac{\\partial F}{\\partial x_{n-2}}=0$ imples $t=0$; and if $t=0$, then $\\frac{\\partial F}{\\partial x_2}=0$ implies $s=0$. Thus, polynomials $F$ of this form give hypersurfaces smooth along $C$.\n\nNow, look at the map $\\psi_F$. One way to obtain a $\\psi_F$ with clear relations between its entries is to choose increasing exponents for $t$. This way, we get a relation between every two consecutive entries by multiplying the first by a power of $t$ and the second by a power of $s$. We will also need zero entries to get the terms $\\mathcal{O}(n+2)$ of $E$. So, let $0\\le z\\le n-3$ and  $0\\le \\beta_2\\le \\beta _3\\le \\cdots \\le \\beta_{n-z-2}\\le n(d-2)+z+1$ be integers, and let $F_i$ be such that\n\\[ F_i|_C = s^{n(d-2)-\\beta_i}t^{\\beta_i} \\text{ for } 2\\le i\\le n-z-2, \\]\nand\n\\[ F_i|_C = 0 \\text{ for } n-z-1\\le i\\le n-2. \\]\n\nNote that such $F_i$ exist by Lemma \\ref{projectivelynormal}. Then, $\\psi_F$ has increasing powers of $t$ and $z$ zero entries:\n\\begin{multline*}\n \\psi_F = ( s^{n-2}(s^{n(d-2)}), \\ (s^{n-3}t)(s^{n(d-2)-\\beta_2}t^{\\beta_2}), \\cdots , (s^{z+1}t^{n-z-3})(s^{n(d-2)-\\beta_{n-z-2}}t^{\\beta_{n-z-2}}), \\ 0, \\ 0, \\cdots \\\\\n \\cdots 0, \\ t^{n-2}(t^{n(d-2)}) ).\n\\end{multline*}\n\nIts columns $C_1, C_2, \\cdots , C_{n-1}$ satisfy the column relations\n\\begin{align*}\n   & t^{\\beta_2+1}C_1 - s^{\\beta_2+1}C_2 = 0, \\\\\n    & t^{\\beta_{i+1}-\\beta_i+1}C_i - s^{\\beta_{i+1}-\\beta_i+1}C_{i-1} = 0, \\ \\ 2\\le i\\le n-z-3, \\\\\n    & C_j = 0, \\ \\ n-z-1\\le j\\le n-2, \\\\\n    & t^{n(d-2)+z+1-\\beta_{n-z-2}}C_{n-z-2} - s^{n(d-2)+z+1-\\beta_{n-z-2}}C_{n-1} = 0.\n\\end{align*}\n\nTo simplify notation and make the final result simpler to parse, we write $\\gamma_i = \\beta_{i+1} - \\beta_i$ for $2\\le i\\le n-z-3$, $\\gamma_1 = \\beta_2$ and $\\gamma_{n-z-2} = \\beta_{n-z-2}$. The conditions for $\\beta_i$ translate into the conditions $\\gamma_i\\ge 0$ and $\\gamma_{n-z-2} = \\sum_{i=1}^{n-z-3}\\gamma_i\\le n(d-2)+z+1$. Rewriting the column relations:\n\\begin{align*}\n   & t^{\\gamma_1+1}C_1 - s^{\\gamma_1+1}C_2 = 0, \\\\\n    & t^{\\gamma_i+1}C_i - s^{\\gamma_i+1}C_{i-1} = 0, \\ \\ 2\\le i\\le n-z-3, \\\\\n    & C_j = 0, \\ \\ n-z-1\\le j\\le n-2, \\\\\n    & t^{n(d-2)+z+1-\\gamma_{n-z-2}}C_{n-z-2} - s^{n(d-2)+z+1-\\gamma_{n-z-2}}C_{n-1} = 0.\n\\end{align*}\n\nWe define the matrix $K$ whose columns give the relations above:\n\\[ K = \\begin{pmatrix}\nt^{\\gamma_1+1} & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 \\\\\n-s^{\\gamma_1+1} & t^{\\gamma_2+1} & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 \\\\\n0 & -s^{\\gamma_2+1} & t^{\\gamma_3+1} & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 \\\\\n\\vdots &  \\vdots & \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots\\\\\n0 &  0  & 0 & \\cdots & t^{\\gamma_{n-z-3}+1} & 0 & 0 & \\cdots & 0 & 0\\\\\n0 &  0  & 0 & \\cdots & -s^{\\gamma_{n-z-3}+1} & 0 & 0 & \\cdots & 0 & t^{n(d-2)+z+1-\\gamma_{n-z-2}}\\\\\n0 &  0  & 0 & \\cdots & 0 & 1 & 0 &\\cdots & 0 & 0\\\\\n0 &  0  & 0 & \\cdots & 0 & 0 & 1 & \\cdots & 0 & 0\\\\\n\\vdots &  \\vdots & \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots \\\\\n0 &  0  & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 1 & 0\\\\\n0 &  0  & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & -s^{n(d-2)+z+1-\\gamma_{n-z-2}}\n\\end{pmatrix}\\]\n\nThen $K$ defines a map\n\\[ K: \\left (\\bigoplus_{i=1}^{n-z-3} \\mathcal{O}(n+1-\\gamma_i)\\right )\\oplus \\mathcal{O}(n+2)^z\\oplus \\mathcal{O}(n+1 - n(d-2)-z+\\gamma_{n-z-2})\\to \\mathcal{O}(n+2)^{n-1}. \\]\n\nThe column relations imply that the image of $K$ is contained in the kernel $N_{C/X}$ of $\\psi_F$. Since $K$ has maximal rank, its image is a subbundle of $N_{C/X}$. As both $N_{C/X}$ and the subbundle have rank $n-2$, it follows that\n\\[ N_{C/X}\\cong \\mathcal{O}(n+2)^z\\oplus \\left (\\bigoplus_{i=1}^{n-z-3} \\mathcal{O}(n+1-\\gamma_i)\\right )\\oplus \\mathcal{O}(n+1 - n(d-2)-z+\\gamma_{n-z-2}).\\]\n\nThus, we get splitting types of $N_{C/X}$ with exactly $z$ terms $\\mathcal{O}(n+2)$. By appropriately choosing the $\\gamma_i\\ge 0$ with $\\gamma_{n-z-2} = \\sum_{i=1}^{n-z-3}\\gamma_i\\le n(d-2)+z+1$, or equivalently, by choosing the $\\beta_i$ in increasing order, we get all possible choices for the summands of smaller degrees. Note the last term is determined by the previous ones and the degree of $N_{C/X}$. Finally, by varying the number of zero entries $z$ between $0$ and $n-3$, we obtain all splitting types $E$ for $N_{C/X}$.\n\n\\medskip\n\nNow, we investigate the case $e<n$. Similarly, consider polynomials $F$ of the form $F = \\sum _{i=1}^{e-1}F_iQ_{i,i+1} + \\sum _{j=e+1}^nG_jx_j$ with $F_1 = x_0^{d-2}$ and $G_n = x_e^{d-1}$, so that $N_{C/X}$ is the kernel of maps $\\psi_F:\\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}\\to \\mathcal{O}(de)$ of the form\n\\[ \\psi_F = \\left ( s^{e-2}(s^{e(d-2)}), \\ (s^{e-3}t)F_2|_C, \\cdots , t^{e-2}F_{e-1}|_C \\ ; \\ G_{e+1}|_C, \\ \\cdots , G_{n-1}|_C, \\ t^{e(d-1)} \\right ). \\]\n\nThe approach in this case is similar. The main difference is that $N_{C/\\P^n}$ splits into two summands of different degrees: $\\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}$. We separate $\\psi_F$ into two parts as above, corresponding to the $F_i$ and $G_j$. We will choose the zero entries all in the first part so we can get the $\\mathcal{O}(e+2)$ terms of $E$.\n\nFirst, we show that these $F$ define hypersurfaces smooth along the curve. Taking partial derivatives $\\frac{\\partial F}{\\partial x_i}$ and restricting to the curve (letting $Q_{ij} = 0$ and $x_l = 0$, $l\\ge e+1$), we obtain\n\\[ \\frac{\\partial F}{\\partial x_2} = -x_0^{d-1} + 2x_2F_2 - F_3x_4 \\ \\ \\ \\text{ and } \\ \\ \\ \\frac{\\partial F}{\\partial x_n} = x_e^{d-1}. \\]\n\nSo, if $F$ is singular at a point $P=(s^e : s^{e-1}t : \\cdots : t^e : 0 : \\cdots : 0)$ on $C$ then $x_e^{d-1} = 0$ implies $t=0$, thus $x_2 = x_4 =0$, and the first equation gives $x_0^{d-1} = 0$, so $s=0$. Hence, all $F$ of this form are smooth along $C$.\n\nAs in the case $e=n$, we choose maps $\\psi_F$ with $z$ zero entries and increasing exponents of $t$, so to get relations between every two consecutive entries. This time, consider integers\n\\[ 0\\le z\\le e-2, \\ \\ 0\\le \\beta_2\\le \\beta_3\\le \\cdots \\le \\beta_{e-1-z} \\ \\text{ and } \\ \\beta_{e-1-z}-z+e\\le \\beta_e\\le \\cdots \\le \\beta_{n-2}\\le e(d-1), \\]\nand let $F_i$ and $G_j$ be such that\n\\begin{align*}\n& F_i|_C = s^{e(d-2)-\\beta_i}t^{\\beta_i} \\text{ for } 2\\le i\\le e-1-z,\\\\\n& F_i|_C = 0 \\text{ for } e-z\\le i\\le e-1 \\text{ and }\\\\\n& G_j|_C = s^{e(d-1)-\\beta_{j-1}}t^{\\beta_{j-1}} \\text{ for } e+1\\le j\\le n-1.\n\\end{align*}\nso the map $\\psi_F$ has the form\n\\begin{multline*}\n\\psi_F =  ( s^{e-2}(s^{e(d-2)}), \\ \\ (s^{e-3}t)(s^{e(d-2)-\\beta_2}t^{\\beta_2}), \\ \\cdots , \\ (s^zt^{e-2-z})(s^{e(d-2)-\\beta_{e-1-z}}t^{\\beta_{e-1-z}}), \\ \\ 0, \\ \\cdots , \\ 0 ; \\\\\n(s^{e(d-1)-\\beta_e}t^{\\beta_e}), \\ \\ (s^{e(d-1)-\\beta_{e+1}}t^{\\beta_{e+1}}), \\ \\cdots , \\ (s^{e(d-1)-\\beta_{n-2}}t^{\\beta_{n-2}}), \\ \\ t^{e(d-1)} ),\n\\end{multline*}\n\nWe obtain the following relations between the columns $C_1, \\cdots , C_{e-1}; C_e, \\cdots C_{n-1}$:\n\\begin{align*}\n    & t^{\\beta_2+1}C_1 - s^{\\beta_2+1}C_2 = 0\\\\\n    & t^{\\beta_{i+1}-\\beta_i+1}C_i - s^{\\beta_{i+1}-\\beta_i+1}C_{i+1} = 0, \\text{ for } 2\\le i\\le e-2-z\\\\\n    & C_i = 0, \\text{ for } e-z\\le e-1\\\\\n    & t^{\\beta_e-\\beta_{e-1-z}-e+2+z}C_{e-2-z} - s^{\\beta_e-\\beta_{e-1-z}+z-e}C_{e} = 0\\\\\n    & t^{\\beta_{j+1}-\\beta_j}C_j - s^{\\beta_{j+1}-\\beta_j}C_{j+1} = 0, \\text{ for } e\\le j\\le n-3\\\\\n    & t^{e(d+1)-\\beta_{n-2}}C_{n-2}-s^{e(d+1)-\\beta_{n-2}}C_{n-1} = 0.\n\\end{align*}\n\nAs before, rename $\\gamma_i = \\beta_{i+1} - \\beta_i$, $\\gamma_1 = \\beta_2$, $\\gamma_{e-1} = \\beta_e - \\beta_{e-1-z} - e + z$ and $\\gamma_{n-2} = \\beta_{n-2}$. These satisfy the conditions $\\gamma_i\\ge 0$ for all $i$ and $\\gamma_{n-2} = \\sum_{i=1}^{n-2-z}\\gamma_i + \\sum_{j=e-1}^{n-3}\\gamma_j\\le e(d-1)$. Then the column relations become\n\\begin{align*}\n    & t^{\\gamma_2+1}C_1 - s^{\\gamma_2+1}C_2 = 0\\\\\n    & t^{\\gamma_i+1}C_i - s^{\\gamma_i+1}C_{i+1} = 0, \\text{ for } 2\\le i\\le e-2-z\\\\\n    & C_i = 0, \\text{ for } e-1-z\\le e-1\\\\\n    & t^{\\gamma_{e-1}+2}C_{e-2-z} - s^{\\gamma_{e-1}}C_{e} = 0\\\\\n    & t^{\\gamma_j}C_j - s^{\\gamma_j}C_{j+1} = 0, \\text{ for } e\\le j\\le n-3\\\\\n    & t^{e(d+1)-\\gamma_{n-2}}C_{n-2}-s^{e(d+1)-\\gamma_{n-2}}C_{n-1} = 0.\n\\end{align*}\n\nThey induce the matrix\n\n\\[ K = \\begin{pmatrix}\nt^{\\gamma_1+1} & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 \\\\\n-s^{\\gamma_1+1} & t^{\\gamma_2+1} & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 \\\\\n0 & -s^{\\gamma_2+1} & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0  & \\cdots & 0 & 0 \\\\\n\\vdots &  \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots \\\\\n0 &  0  & \\cdots & t^{\\gamma_{e-z-2}+1} & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0\\\\\n0 &  0  & \\cdots & -s^{\\gamma_{e-z-2}+1} & 0 & 0 & \\cdots & 0 & t^{\\gamma_{e-1}+2} & 0 & \\cdots & 0 & 0\\\\\n0 &  0  & \\cdots & 0 & 1 & 0 &\\cdots & 0 & 0 & 0 & \\cdots & 0 & 0\\\\\n0 &  0  & \\cdots & 0 & 0 & 1 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0\\\\\n\\vdots &  \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots  & \\vdots & \\ddots & \\vdots & \\vdots\\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 1 & 0 & 0 & \\cdots & 0 & 0\\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 0 & -s^{\\gamma_{e-1}} & t^{\\gamma_e} & \\cdots & 0 & 0 \\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & -s^{\\gamma_e} & \\cdots & 0 & 0 \\\\\n\\vdots &  \\vdots & \\ddots & \\vdots & \\vdots & \\vdots & \\ddots & \\vdots & \\vdots  & \\vdots & \\ddots & \\vdots & \\vdots\\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & t^{\\gamma_{n-3}} & 0\\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & -s^{\\gamma_{n-3}} & t^{e(d+1)-\\gamma_{n-2}}\\\\\n0 &  0  & \\cdots & 0 & 0 & 0 & \\cdots & 0 & 0 & 0 & \\cdots & 0 & -s^{e(d+1)-\\gamma_{n-2}}\n\\end{pmatrix}. \\]\n\n\\medskip\n\nAnd $K$ defines an injective map\n\\[ K: \\left( \\bigoplus _{i=1}^{e-2-z}\\mathcal{O}(e+1-\\gamma_i) \\right )\n\\oplus \\mathcal{O}(e+2)^{z}\\oplus \\left ( \\bigoplus _{j=e-1}^{n-3}\\mathcal{O}(e-\\gamma_j) \\right ) \\oplus \\mathcal{O}(e-e(d+1)+\\gamma_{n-2})\\to \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}. \\]\n\nAs before, the column relations imply that $K$ factors through the kernel of $\\psi_F$, and consequently, the kernel is\n\\[ N_{C/X}\\cong \\mathcal{O}(e+2)^{z}\\oplus \\left( \\bigoplus _{i=1}^{e-2-z}\\mathcal{O}(e+1-\\gamma_i) \\right )\n\\oplus \\left ( \\bigoplus _{j=e-1}^{n-3}\\mathcal{O}(e-\\gamma_j) \\right ) \\oplus \\mathcal{O}(e-e(d+1)+\\gamma_{n-2}). \\]\n\nThis is a splitting type with exactly $z$ terms $\\mathcal{O}(e+2)$, $e-2-z$ terms of degree less or equal than $e+1$, and the rest of them of degree less or equal than $e$. By varying the $\\gamma_i$ with $\\gamma_i\\ge 0$ and $\\gamma_{n-2} = \\sum_{i=1}^{n-2-z}\\gamma_i + \\sum_{j=e-1}^{n-3}\\gamma_j\\le e(d-1)$, we get all splitting types $E$ of this form. Finally, by varying the number of zero entries $z$ between $0$ and $n-2$ we obtain all possible splitting types $E$ for $N_{C/X}$.\n\\end{proof}\n\n\\begin{corollary}\\label{smoothexample}\n(char $K=0$) Let $d\\ge 4$, and assume the base field $K$ has characteristic $0$. For all splitting types $E$ as in the theorem, there exists a smooth hypersurface $X$ of degree $d$ containing the curve $C$ with normal bundle $N_{C/X}\\cong E$.\n\\end{corollary}\n\\begin{proof} Let $\\psi_F\\in \\Hom (N_{C/\\P^n},\\mathcal{O}(de))$ be the map with $N_{C/X}\\cong E$, induced by the polynomial $F$ obtained in the theorem. From the exact sequence\n\\[ 0\\longrightarrow H^0(\\mathcal{I}_C^2(d))\\longrightarrow H^0(\\mathcal{I}_C(d))\\overset{\\phi}{\\longrightarrow} \\Hom (N_{C/\\P^n},\\mathcal{O}(de)), \\]\nwe have that the space of polynomials that induce the same $\\psi_F$ is the coset \n\\[ F+H^0(\\mathcal{I}_C^2(d)) = \\{ F+G \\ | \\ G\\in H^0(\\mathcal{I}_C^2(d)) \\}. \\]\n\nIf $d\\ge 4$, then multiples of $Q_{ij}^2$ are in $H^0(\\mathcal{I}_C^2(d))$, so the base locus of $F + H^0(\\mathcal{I}_C^2(d))$ is the curve $C$. Thus, by Bertini's theorem, the general member of $F+H^0(\\mathcal{I}_C^2(d))$ is smooth away from $C$. As $F$ is smooth along $C$, the general member of $F+H^0(\\mathcal{I}_C^2(d))$ is smooth.\n\\end{proof}\n\n\\subsection{Dimension Count}\n\nTheorem \\ref{existence} shows that for all possible splitting types $E$ of $N_{C/X}$ we can compute an hypersurface $X$ for which $N_{C/X}\\cong E$. A natural question that arises is if we can compute the dimension of the space of such hypersurfaces, and if it is the expected one. \n\nLet $E_{\\vec{a}} = \\bigoplus _{i=1}^{n-2}\\mathcal{O}(a_i)$ be a possible splitting type. The numerical conditions in Theorem \\ref{existence} translate to $a_i\\le e+2$ for $1\\le i\\le e-2$, $a_j\\le e$ for $e-1\\le j\\le n-2$, and $\\sum_{i=1}^{n-2}a_i = e(n-d+1)-2$.\n\nIn the introduction, we defined the spaces\n\\[ \\Sigma = \\{ F \\ | \\ X \\text{ is a degree } d \\text{ hypersurface smooth along } C \\}\\subset H^0(\\mathcal{O}_{\\P^n}(d)) \\]\nand\n\\[ \\Sigma_{\\vec{a}} = \\{ F\\in \\Sigma \\ | \\ N_{C/X}\\cong E_{\\vec{a}} \\}\\subset \\Sigma , \\]\nand observed that the expected codimension of $\\Sigma_{\\vec{a}}$ in $\\Sigma$ is $h^1(\\End(E_{\\vec{a}}))$.\n\nWe start by computing this codimension at the level of homomorphisms. For that, define\n\\[ \\Phi_{\\vec{a}} = \\{ M\\in \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de)) \\ | \\ \\ker M\\cong E_{\\vec{a}} \\}\\subset \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de)). \\]\n\n\\begin{proposition}\\label{lemmaHom} The codimension of $\\Phi_{\\vec{a}}$ in $\\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))$ is\n\\[ h^1(\\End (E_{\\vec{a}})) - h^1(\\sHom(E_{\\vec{a}}, \\mathcal{O}(de))). \\]\n\\end{proposition}\n\\begin{proof}\nWe first compute the dimension of $\\Phi_{\\vec{a}}$. Every $M$ in $\\Phi_{\\vec{a}}$ comes from an injection $\\theta: E_{\\vec{a}}\\hookrightarrow N_{C/\\P^n}$, and every injection defines an $M$, since its cokernel is a line bundle of degree $de$. By our numerical conditions on $\\vec{a}$, $\\sHom (E_{\\vec{a}}, N_{C/\\P^n})$ is globally generated, so the dimension of injections $\\theta$ is $h^0(\\sHom(E_{\\vec{a}}, N_{C/\\P^n}))$. However, each $M$ might come from different injections. To count these, consider the set of pairs of homomorphisms\n\\[ B = \\{ (\\theta, M)\\mid E_{\\vec{a}}\\overset{\\theta}{\\hookrightarrow} N_{C/\\P^n}\\overset{M}{\\twoheadrightarrow} \\mathcal{O}(de) \\}\\subset \\Hom (E_{\\vec{a}}, N_{C/\\P^n})\\times \\Hom (N_{C/\\P^n}, \\mathcal{O}(de)) \\]\ntogether with the projections $\\pi_1$ and $\\pi_2$:\n\\[ \\begin{tikzcd}\n& B \\arrow[ld, \"\\pi_1\"'] \\arrow[rd, \"\\pi_2\"] & \\\\\n\\Hom (E_{\\vec{a}}, N_{C/\\P^n}) & & \\Hom (N_{C/\\P^n}, \\mathcal{O}(de))\n\\end{tikzcd} \\]\n\n\nA fiber $\\pi^{-1}(\\theta)$ of the map $\\pi_1$ correspond to the $M$'s induced by $\\theta$ up to a choice of basis for $\\mathcal{O}(de)$, thus, the fibers of $\\pi_1$ are isomorphic to $\\Aut(\\mathcal{O}(de))$. The image of $\\pi_1$ is the open set of injections $\\theta \\in \\Hom (E_{\\vec{a}}, N_{C/\\P^n})$. Hence, $\\dim B = h^0(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) + h^0(\\End(\\mathcal{O}(de)))$.\n\nOn the other hand, the image of $\\pi_2$ is exactly the set $\\Phi_{\\vec{a}}$, and the fibers of $\\pi_2$ are isomorphic to $\\Aut (E_{\\vec{a}})$. Thus,\n\\[ \\dim \\Phi_{\\vec{a}} = h^0(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) + h^0(\\End(\\mathcal{O}(de))) - h^0(\\End(E_{\\vec{a}})). \\]\n\nTherefore, its codimension in $\\Hom (N_{C/\\P^n}, \\mathcal{O}(de))$ is\n\\[ \\codim \\Phi_{\\vec{a}} = h^0(\\sHom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))) - h^0(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) + h^0(\\End(E_{\\vec{a}})) - h^0(\\End(\\mathcal{O}(de))). \\]\n\nNow, applying the functors $\\sHom (E_{\\vec{a}}, - )$ and $\\sHom (-, \\mathcal{O}(de))$ to the sequence\n\\[ 0\\longrightarrow E_{\\vec{a}}\\longrightarrow N_{C/\\P^n}\\longrightarrow \\mathcal{O}(de)\\longrightarrow 0 \\]\nwe obtain the long exact sequences\n\\begin{align*}\n     0\\longrightarrow & \\sHom(E_{\\vec{a}}, E_{\\vec{a}})\\longrightarrow \\sHom(E_{\\vec{a}}, N_{C/\\P^n})\\longrightarrow \\sHom(E_{\\vec{a}}, \\mathcal{O}(de))\\longrightarrow \\\\\n     \\longrightarrow & \\sExt^1(E_{\\vec{a}}, E_{\\vec{a}})\\longrightarrow \\sExt^1(E_{\\vec{a}}, N_{C/\\P^n})\\longrightarrow \\sExt^1(E_{\\vec{a}}, \\mathcal{O}(de)) = 0\n\\end{align*}\nand\n\\[ 0\\longrightarrow \\sHom(\\mathcal{O}(de), \\mathcal{O}(de))\\longrightarrow \\sHom(N_{C/\\P^n}, \\mathcal{O}(de))\\longrightarrow \\sHom(E_{\\vec{a}}, \\mathcal{O}(de))\\longrightarrow 0. \\]\n\nObserve that $\\sExt^1(E_{\\vec{a}}, \\mathcal{O}(de)) = 0$ since $a_i\\le e+2\\le de$ for all $1\\le i\\le n-2$.\nComputing the alternating sum on dimensions of the first sequence, and substituting it on $\\codim \\Phi_{\\vec{a}}$, we get\n\\begin{multline*}\n\\codim \\Phi_{\\vec{a}} = h^0(\\sHom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))) - h^0(\\sHom(E_{\\vec{a}}, \\mathcal{O}(de)) + h^1(\\End(E_{\\vec{a}}))\\\\\n- h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) - h^0(\\End(\\mathcal{O}(de))).\n\\end{multline*}\n\nAnd from the second sequence it follows that\n\\begin{align*}\n    \\codim \\Phi_{\\vec{a}} = & h^0(\\End(\\mathcal{O}(de))) + h^0(\\sHom(E_{\\vec{a}}, \\mathcal{O}(de))) - h^0(\\sHom(E_{\\vec{a}}, \\mathcal{O}(de)) \\\\\n    & \\hspace{2.5cm}  + h^1(\\End(E_{\\vec{a}}) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) - h^0(\\End(\\mathcal{O}(de)))\\\\\n    = & h^1(\\End(E_{\\vec{a}}) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})).\n\\end{align*}\n\\end{proof}\n\n\\begin{corollary}\\label{corhom} If $e=n$ or $a_i\\le e+1$ for all $i$, then the codimension of $\\Phi_{\\vec{a}}$ in $\\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(de))$ is $h^1(\\End(E_{\\vec{a}}))$.\n\\end{corollary}\n\\begin{proof}\nIn this case, since $N_{C/\\P^n}\\cong \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}$, we have $h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) = 0$.\n\\end{proof}\n\nWe can compute the dimension of the balanced splitting type, recovering a particular case of \\cite{CR}, Corollary 2.2.\n\n\\begin{corollary}\\label{phiissurjective} (\\cite{CR}, Corollary 2.2) \nThe kernel of the general $M\\in \\Hom (N_{C/\\P^n}, \\mathcal{O}(de))$ is balanced.\n\\end{corollary}\n\\begin{proof}\nFor $E_{\\vec{a}}$ the balanced splitting type we have $a_i\\le \\left \\lceil \\frac{e(n-d+1)-2}{n-2} \\right \\rceil \\le e+1$ and $|a_i-a_j|\\le 1$ for all $i,j$, thus\n\\[ \\codim \\Phi_{\\vec{a}} = h^1(\\End(E_{\\vec{a}})) = 0. \\]\n\nHence $\\Phi_{\\vec{a}}$ has dimension equal to $\\dim \\Hom (N_{C/\\P^n}, \\mathcal{O}(de))$.\n\\end{proof}\n\n\\begin{theorem}\\label{teoremasigmaa} The locus $\\Sigma_{\\vec{a}}$ is irreducible and smooth of codimension $h^1(\\End(E_{\\vec{a}})) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n}))$ in $\\Sigma$.\n\\end{theorem}\n\\begin{proof}\nLet $\\overline{\\Sigma_{\\vec{a}}}$ be the space of all polynomials $F\\in H^0(\\mathcal{I}_C(d))$, not necessarily smooth along $C$, whose kernel of $\\psi_F$ has splitting type $E_{\\vec{a}}$. Consider the map $\\beta :\\overline{\\Sigma_{\\vec{a}}} \\to \\Phi_{\\vec{a}}$, $\\beta (F) = \\psi _F$. By \\cite{CR} Theorem 3.1, the map $\\phi : H^0(\\mathcal{I}_C(d))\\to \\Hom (N_{C/\\P^n},\\mathcal{O}(de))$ is surjective. It follows that $\\beta$ is also surjective. Note its image $\\Phi_{\\vec{a}}$ is irreducible, since it is the image of $\\pi_2$ from the last proposition. In addition, by the short exact sequence \n\\[ 0\\longrightarrow H^0(\\mathcal{I}_{C/\\P^n}^2(d))\\longrightarrow H^0(\\mathcal{I}_{C/\\P^n}(d))\\overset{\\phi}{\\longrightarrow} H^0(\\Hom (N_{C/\\P^n}, \\mathcal{O}(de)))\\longrightarrow 0, \\]\nthe fiber $\\beta ^{-1}(M)$ is isomorphic to the linear system $H^0(\\mathcal{I}_C^2(d))$. Thus, $\\beta ^{-1}(M)$ is smooth and irreducible of dimension $h^0(\\mathcal{I}^2_C(d))$. Then, by Proposition \\ref{lemmaHom}, it follows that $\\overline{\\Sigma_{\\vec{a}}}$ is smooth and irreducible of dimension\n\\[ h^0(\\sHom(N_{C/\\P^n}, \\mathcal{O}(de))) - h^1(\\End(E_{\\vec{a}})) + h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) + h^0(\\mathcal{I}^2_C(d)). \\]\n\nSince $F$ being smooth along $C$ is an open condition in $\\overline{\\Sigma_{\\vec{a}}}$, and $\\Sigma_{\\vec{a}}$ is not empty by Theorem \\ref{existence}, it follows that $\\Sigma_{\\vec{a}}$ is an open dense subset of $\\overline{\\Sigma_{\\vec{a}}}$. Therefore, $\\Sigma_{\\vec{a}}$ is irreducible and smooth of the same dimension.\n\nThe dimension of $\\Sigma$ is $h^0(\\mathcal{I}_C(d))$, and by the sequence above,\n\\[ h^0(\\mathcal{I}^2_C(d)) = h^0(\\mathcal{I}_C(d)) - h^0(\\sHom(N_{C/\\P^n}, \\mathcal{O}(de))). \\]\n\nHence, the codimension of $\\Sigma _{\\vec{a}}$ in $\\Sigma$ is\n\\[ \\codim (\\Sigma_{\\vec{a}}\\subset \\Sigma) = h^1(\\End(E_{\\vec{a}})) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})). \\]\n\\end{proof}\n\nBy Corollary \\ref{smoothexample}, when $d\\ge4$ the general hypersurface of $\\Sigma_{\\vec{a}}$ is smooth.\n\n\\begin{corollary}\n(char $K=0$) Let $d\\ge 4$, and assume the base field $K$ has characteristic $0$. Let $S\\Sigma_{\\vec{a}}$ be the subspace of $\\Sigma_{\\vec{a}}$ of polynomials $F$ with $X$ smooth. Then $S\\Sigma_{\\vec{a}}$ is irreducible and smooth of codimension $h^1(\\End(E_{\\vec{a}})) - h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n}))$ in $\\Sigma$.\n\\end{corollary}\n\nAs in Corollary \\ref{corhom}, we get the expected codimension when $e=n$ or all $a_i$ are smaller than $e+2$.\n\n\\begin{corollary} If $e=n$ or $a_i\\le e+1$ for all $i$, then the codimension of $\\Sigma_{\\vec{a}}$ in $\\Sigma$ is the expected $h^1(\\End(E_{\\vec{a}}))$.\n\\end{corollary}\n\nWhen $e < n$ and there exist terms $a_i = e+2$, the expected and the actual codimension differ. We can compute this difference as follows.\n\n\\begin{corollary}\\label{difference} Let $z = |\\{ i \\ | \\ a_i = e+2 \\}|$ be the number of terms $\\mathcal{O}(e+2)$ in the splitting type $E_{\\vec{a}}$. Then the codimension of $\\Sigma_{\\vec{a}}$ in $\\Sigma$ is $h^1(\\End(E_{\\vec{a}})) - (n-e)z$.\n\\end{corollary}\n\\begin{proof}\nBy Theorem \\ref{teoremasigmaa} and Serre duality for $\\P^1$,\n\\begin{align*}\n h^1(\\sHom(E_{\\vec{a}}, N_{C/\\P^n})) & = h^1(\\sHom(E_{\\vec{a}}, \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}))  \\\\ \n    & = h^0(\\sHom(\\mathcal{O}, (\\mathcal{O}(-e-2)^{e-1}\\oplus \\mathcal{O}(-e)^{n-e})\\otimes (\\oplus_{i=1}^{n-2} \\mathcal{O}(a_i))\\otimes \\mathcal{O}(-2)))  \\\\\n    & = \\sum_{i=1}^{n-2} h^0(\\P^1, \\mathcal{O}(-e-4+a_i)^{e-1}\\oplus \\mathcal{O}(-e-2+a_i)^{n-e}) = (n-e)z.\n\\end{align*}\n\\end{proof}\n\n\nIn particular, the difference between the actual and the expected codimension can get arbitrarily large as $n$ grows.\n\n\n\\section{Quadric Hypersurfaces}\n\nIn this section, we study the case $d=2$. Let $X = V(F)$ be a degree 2 hypersurface that contains the rational normal curve $C$ of degree $e$, and consider the exact sequence of the map $\\phi$ described in Section 2:\n\\[ 0\\longrightarrow H^0(\\mathcal{I}^2_C(2))\\longrightarrow H^0(\\mathcal{I}_C(2))\\overset{\\phi}{\\longrightarrow} \\Hom (N_{C/\\P^n}, \\mathcal{O}_{\\P^1}(2e)). \\]\n\nUnlike when $d\\ge 3$, the map $\\phi$ is not surjective. Thus, the cokernel of an injection $E_{\\vec{a}}\\hookrightarrow N_{C/\\P^n}$ may not be in the image of $\\phi$, so we cannot repeat the arguments from Proposition \\ref{lemmaHom} to compute the dimension of $\\Sigma_{\\vec{a}}$. Nevertheless, we can compute the dimension of the image of $\\phi$.\n\n\\begin{proposition} The dimension of $H^0(\\mathcal{I}_C^2(2))$ is $\\frac{(n-e)(n-e+1)}{2}$. In particular, $\\phi$ is injective when $e=n$.\n\n\\end{proposition}\n\\begin{proof}\nIf $F\\in H^0(\\mathcal{I}_C^2(2))$ then $F$ is double along $C$. Since $F$ is a quadric, and the singular locus of a quadric is a linear space, $F$ must also be singular on all the $\\P^e$ span by $C$. Thus, $F$ must be a combination of the generators $x_ix_j$, $e+1\\le i,j\\le n$ of $\\P^e$. That is,\n    \\[ F = \\sum _{e+1\\le i\\le j\\le n} \\lambda_{ij}x_ix_j, \\ \\ \\lambda_{ij}\\in K. \\]\n    \n    So we got to choose $\\frac{(n-e)(n-e+1)}{2}$ coefficients to define $F$. Thus, $h^0(\\mathcal{I}_C^2(2)) = \\frac{(n-e)(n-e+1)}{2}$.\n\\end{proof}\n\n\\begin{corollary} The image of $\\phi$ has dimension $\\frac{2ne + 2n -3e - e^2}{2}$.\n\\end{corollary}\n\\begin{proof}\nFrom the exact sequence,\n\\[ \\dim (\\mathrm{im} \\ \\phi ) = h^0(\\mathcal{I}_C(2)) - h^0(\\mathcal{I}^2_C(2)) = \\dbinom{n+2}{2} - (2e+1) - \\frac{(n-e)(n-e+1)}{2} = \\frac{2ne + 2n -3e - e^2}{2}. \\]\n\\end{proof}\n\n\\subsection{Splitting Types}\n\nAs in the case $d\\ge 3$, we start by showing that we can obtain examples of quadrics smooth along $C$ for every possible splitting type of $N_{C/X}$.\n\nThe normal bundle sequence in the case $d=2$ is\n\\[ 0\\longrightarrow N_{C/X}\\longrightarrow \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}\\overset{\\psi_F}{\\longrightarrow}\\mathcal{O}(2e)\\longrightarrow 0 \\]\nand $\\deg N_{C/X} = e(n-1)-2$. Then, any splitting type for $N_{C/X}$ must have the form of a direct sum $\\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right )$ with $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i+\\sum_{j=e+1}^nb_j = e-2$.\n\n\\begin{theorem}\\label{quadricprop1}\nFor any given splitting type \n\\[ E = \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right ), \\] with $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^nb_j = e-2$, we produce a quadric $X=V(F)$, smooth along $C$, for which $N_{C/X}\\cong E$.\n\nMore explicitly, for $e=n$, rearrange the $a_i$ in increasing order $a_1\\le \\cdots \\le a_{n-2}$, and let $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $i=1, \\cdots , n-2$. Then the quadric $X$ given by the polynomial\n\\[ F = \\sum_{i=0}^{n-2}Q_{\\beta_i+1,\\beta_i+2} =  Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + Q_{\\beta_2 + 1, \\beta_2 + 2} \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}, \\]\nwhere $Q_{ij} = x_ix_{j-1}-x_{i-1}x_j$, has normal bundle $N_{C/X}\\cong E$.\n\nFor $e<n$, let $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $1\\le i\\le e-2$. Let also $\\gamma_n = e$ and $\\gamma_{j} = e - b_n - b_{n-1} - \\cdots - b_{j+1}$ for $e+1\\le j\\le n-1$. Then a quadric $X$ such that $N_{C/X}\\cong E$ is given by\n\\begin{align*}\n    F = & \\sum_{i=0}^{e-2}Q_{\\beta_i+1,\\beta_i+2} + \\sum_{j=e+1}^n x_{\\gamma_j}x_j\\\\\n    = & (Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}) + (x_{\\gamma_j}x_{e+1} + \\cdots + x_{\\gamma_{n-1}}x_{n-1} + x_ex_n).\n\\end{align*}\n\\end{theorem}\n\nBefore we proceed to the proof of the theorem, let us use an example to better illustrate the idea of the proof, and to show how we are using the relations between the entries of $\\psi_F$ to compute its kernel.\n\nLet $n=e=5$. Consider polynomials $F$ of the form $F = \\lambda_1Q_{1,2} + \\lambda_2Q_{2,3} + \\lambda_3Q_{3,4} + \\lambda_4Q_{4,5}$. So $F$ induces the map $\\psi_F: \\mathcal{O}(7)^{4}\\to \\mathcal{O}(10)$,\n\\[ \\psi_F = (\\lambda_1s^3, \\ \\lambda_2s^2t, \\ \\lambda_3st^2, \\ \\lambda_4t^3). \\]\n\nWhen $\\lambda_i = 1$, $i=1,\\dots ,4$, $\\psi_F = (s^3, \\ s^2t, \\ st^2, \\ t^3)$ and we get the following degree $1$ relations between consecutive entries of $\\psi_F$, which we call \\textit{column relations} of $\\psi_F$:\n\\begin{align*}\n    & t(s^3) - s(s^2t) = 0\\\\\n    & t(s^2t) - s(st^2) = 0\\\\\n    & t(st^2) - s(t^3) = 0\\\\\n\\end{align*}\nWriting the coefficients of these relations as columns vectors, we get the matrix $K$\n\\[ K = \\begin{pmatrix}\nt & 0 & 0\\\\\n-s & t & 0\\\\\n0 & -s & t\\\\\n0 & 0 & -s\n\\end{pmatrix}, \\]\nwhich defines a map $\\mathcal{O}(6)^3\\overset{K}{\\to} \\mathcal{O}(7)^4$. The column relations imply that the image of $K$ is contained in the kernel of $\\psi_F$. As $K$ has rank 3, the map is injective and it coincides with the kernel. Hence, $N_{C/X}\\cong \\mathcal{O}(6)^3$.\n\nNote that the splitting type of $N_{C/X}$ is determined by the degrees of the column relations of $\\psi_F$. So, if we wish to get $N_{C/X}\\cong \\mathcal{O}(5)\\oplus \\mathcal{O}(6)\\oplus \\mathcal{O}(7)$, we need relations of degree $0$, $1$ and $2$. Consecutive nonzero entries give degree $1$ relations. Similarly, entries separated by one entry give a degree $2$ relation. Then, let $\\lambda_3 = 0$, so $\\psi_F = (s^3, \\ s^2t, \\ 0, \\ t^3)$. It satisfies the column relations\n\\begin{align*}\n    & t(s^3) - s(s^2t) = 0\\\\\n    & t^2(s^2t) - s^2(t^3) = 0\n\\end{align*}\nand similarly, we define the matrix\n\\[ K = \\begin{pmatrix}\nt & 0 & 0\\\\\n-s & 0 & t^2\\\\\n0 & 1 & 0\\\\\n0 & 0 & -s^2\n\\end{pmatrix}, \\]\nand obtain that $N_{C/X}\\cong \\mathcal{O}(6)\\oplus \\mathcal{O}(7)\\oplus \\mathcal{O}(5)\\overset{K}{\\to} \\mathcal{O}(7)^4$ is the kernel of $\\psi_F$.\n\nFinally, the splitting type $\\mathcal{O}(4)\\oplus \\mathcal{O}(7)^2$ must come from a $\\psi_F$ that has a column relation of degree $3$. We get it by letting two consecutive entries of $\\psi_F$ be $0$, that is, let $\\psi_F = (s^3, \\ 0, \\ 0, \\ t^3)$.\n\nIn summary, for every term $\\mathcal{O}(n+2-a_i)$ of the splitting type, we need a corresponding column relation of degree $a_i$. We can get it by letting $a_i-1$ consecutive terms of $\\psi_F$ be $0$. By doing it for all $i$, we get all needed relations.\n\n\\medskip\n\n\\begin{proofofthm4.3}\nLet first $e=n$. In this case, the possible splitting types are $E = \\bigoplus_{i=1}^{n-2}\\mathcal{O}(n+2-a_i)$ with $a_i\\ge 0$ and $\\sum_{i=1}^{n-2}a_i = n-2$. We look at polynomials $F$ of the form $F = \\sum_{i=1}^{n-1} \\lambda_iQ_{i,i+1}$ with $\\lambda_i\\in \\{ 0, 1 \\}$ for all $i$. Then, $F$ induces the map $\\psi_F: \\mathcal{O}(n+2)^{n-1}\\to \\mathcal{O}(2n)$,\n\\[ \\psi_F = (\\lambda_1 s^{n-2}, \\ \\lambda_2 s^{n-3}t, \\cdots , \\lambda_i s^{n-1-i}t^{i-1}, \\cdots , \\lambda_{n-1}t^{n-2}). \\]\n\nAs in the examples above, we want to find a $\\psi_F$ with column relations of degrees $a_i$. We can get it by letting $a_i-1$ consecutive entries of $\\psi_F$ to be $0$ for every $i$. So let $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $i=1, \\cdots , n-2$, and consider the map\n\\[ \\psi_F = (s^{n-2}, \\ 0, \\dots , 0, \\ s^{n-2-\\beta_1}t^{\\beta_1},\\ 0, \\dots , 0,\\  s^{n-2-\\beta_2}t^{\\beta_2}, \\ \\ \\cdots \\ \\ , \\ s^{n-2-\\beta_{n-3}}t^{\\beta_{n-3}}, \\ 0, \\dots , 0 , \\ t^{n-2}). \\]\n\nNotice that $\\psi_F$ is induced by the polynomial\n\\[ F = \\sum_{i=0}^{n-2}Q_{\\beta_i+1,\\beta_i+2} =  Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + Q_{\\beta_2 + 1, \\beta_2 + 2} \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}. \\]\n\nThe entries of $\\psi_F$ satisfy the degree $a_{i+1}$ relations\n\\begin{align*}\n & t^{\\beta_{i+1}-\\beta_i}(s^{n-2-\\beta_i}t^{\\beta_i}) - s^{\\beta_{i+1}-\\beta_i}(s^{n-2-\\beta_{i+1}}t^{\\beta_{i+1}})\\\\\n & = t^{a_{i+1}}(s^{n-2-\\beta_i}t^{\\beta_i}) - s^{a_{i+1}}(s^{n-2-\\beta_{i+1}}t^{\\beta_{i+1}}) \\ \\ \\text{ for } \\ \\ 0\\le i\\le n-3.\n\\end{align*}\n\nSince each relation involves a different pair of columns, they are linearly independent, that is, the matrix $K$ whose columns are the coefficients of the relations has maximal rank. Therefore, $K$ gives the kernel of $\\psi_F$, hence $N_{C/X}\\cong \\bigoplus_{i=1}^{n-2}\\mathcal{O}(n+2-a_i)$.\n\nFinally, we show that these $F$ are smooth along the rational normal curve. Consider the partial derivatives of $F$\n\\[ \\frac{\\partial F}{\\partial x_0} = -x_2, \\ \\ \\frac{\\partial F}{\\partial x_2} = 2\\lambda_2x_2 - x_0 - \\lambda_3x_4, \\ \\ \\frac{\\partial F}{\\partial x_{n-2}} = 2x_{n-2} - x_n - \\lambda_{n-3}x_{n-4}. \\]\nIf F is singular at a point $P = (s^n, s^{n-1}t, \\cdots , t^n)\\in C$, then $\\frac{\\partial F}{\\partial x_0} = 0$ at $P$ implies $s=0$ or $t=0$. If $s=0$, then $\\frac{\\partial F}{\\partial x_{n-2}} = 0$ implies $t=0$; and if $t=0$ then $\\frac{\\partial F}{\\partial x_2} = 0$ implies $s=0$.  Hence, $F$ is smooth along $C$.\n\n\\medskip\n\nNow, let $e<n$. We want to obtain all the possible splitting types $E = \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right )$. Consider polynomials $F$ of the form $F = \\sum_{i=1}^{n-1}\\lambda_iQ_{i,i+1} + \\sum_{j=e+1}^{n}L_jx_j$ with $\\lambda_i\\in \\{0,1\\}$ and $L_j$ linear forms. Then $F$ induces the map $\\psi_F: \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e}\\to \\mathcal{O}(2e)$,\n\\[ \\psi_F = (\\lambda_1 s^{n-2}, \\ \\lambda_2 s^{n-3}t, \\cdots , \\lambda_i s^{n-1-i}t^{i-1}, \\cdots , \\lambda_{n-1}t^{n-2}; \\ L_{e+1}|_C, L_{e+2}|_C, \\cdots , L_n|_C). \\]\n\nWe separate $\\psi_F$ into two parts, one corresponding to the $\\lambda_i$ and the other to the $L_j|_C$. The idea is to use the first part to obtain all the terms $\\mathcal{O}(e+2-a_i)$ of $E$ as in the case $e=n$, and to use the second part for the $\\mathcal{O}(e-b_j)$ terms. \n\nLet $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $1\\le i\\le e-2$. Let also $\\gamma_n = e$ and $\\gamma_{j} = e - b_n - b_{n-1} - \\cdots - b_{j+1}$ for $e+1\\le j\\le n-1$. Consider the map\n\\begin{multline*}\n    \\psi_F = (s^{n-2}, \\ 0, \\dots , 0, \\ s^{n-2-\\beta_1}t^{\\beta_1},\\ 0, \\dots , 0,\\  s^{e-2-\\beta_2}t^{\\beta_2}, \\ \\ \\cdots \\ \\ , \\ s^{e-2-\\beta_{e-2}}t^{\\beta_{e-2}}, \\ 0, \\dots 0 ;\\\\\n     \\ \\ s^{e-\\gamma_{e+1}}t^{\\gamma_{e+1}}, \\ s^{e-\\gamma_{e+2}}t^{\\gamma_{e+2}}, \\dots , s^{e-\\gamma_{n-1}}t^{\\gamma_{n-1}}, \\ t^e ).\n\\end{multline*}\n\nNotice $\\psi_F$ is induced by the polynomial\n\\begin{align*}\n    F = & \\sum_{i=0}^{e-2}Q_{\\beta_i+1,\\beta_i+2} + \\sum_{j=e+1}^n x_{\\gamma_j}x_j\\\\\n    = & (Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}) + (x_{\\gamma_j}x_{e+1} + \\cdots + x_{\\gamma_{n-1}}x_{n-1} + x_ex_n).\n\\end{align*} \n\nWe obtain the following relations between the entries of the first part of $\\psi_F$:\n\\begin{align*}\n & t^{\\beta_{i+1}-\\beta_i}(s^{e-2-\\beta_i}t^{\\beta_i}) - s^{\\beta_{i+1}-\\beta_i}(s^{e-2-\\beta_{i+1}}t^{\\beta_{i+1}}) = 0\\\\\n & \\Leftrightarrow t^{a_{i+1}}(s^{e-2-\\beta_i}t^{\\beta_i}) - s^{a_{i+1}}(s^{e-2-\\beta_{i+1}}t^{\\beta_{i+1}}) = 0 \\ \\ \\text{ for } \\ \\ 0\\le i\\le e-3.\n\\end{align*}\n\nAnd between every two consecutive entries of the second part of $\\psi_F$:\n\\begin{align*}\n & t^{\\gamma_{j+1}-\\gamma_j}(s^{e-\\gamma_j}t^{\\gamma_j}) - s^{\\gamma_{j+1}-\\gamma_j}(s^{e-\\gamma_{j+1}}t^{\\gamma_{j+1}}) = 0\\\\\n & \\Leftrightarrow t^{b_{j+1}}(s^{e-\\gamma_j}t^{\\gamma_j}) - s^{b_{j+1}}(s^{e-\\gamma_{j+1}}t^{\\gamma_{j+1}}) = 0 \\ \\ \\text{ for } \\ \\ e+1\\le j\\le e-1.\n\\end{align*}\n\nNote that $\\gamma_{e+1}-\\beta_{e-2} = e - \\left (\\sum_{i=1}^{e-2} a_i + \\sum_{j=e+2}^n b_j\\right ) = e - (e-2-b_{e+1}) = b_{e+1}+2$. We also get a relation between the first and second part of $\\psi_F$:\n\\begin{align*}\n & t^{\\gamma_{e+1}-\\beta_{e-2}}(s^{e-2-\\beta_{e-2}}t^{\\beta_{e-2}}) - s^{\\gamma_{e+1}-\\beta_{e-2}-2}(s^{e-\\gamma_{e+1}}t^{\\gamma_{e+1}}) = 0 \\\\\n & \\Leftrightarrow t^{b_{e+1}+2}(s^{e-2-\\beta_{e-2}}t^{\\beta_{e-2}}) - s^{b_{e+1}}(s^{e-\\gamma_{e+1}}t^{\\gamma_{e+1}}) = 0.\n\\end{align*}\n\nThe matrix $K$ defined by the coefficients of these relations defines a map\n\\[ K: \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right )\\to \\mathcal{O}(e+2)^{e-1}\\oplus \\mathcal{O}(e)^{n-e} \\]\nthat factors through the kernel $N_{C/X}$ of $\\psi_F$. Since the relations involve different pairs of columns, $K$ has maximum rank, so the map is injective, and it follows that\n\\[ N_{C/X}\\cong \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right ). \\]\n\nTo conclude, we check that $F$ is smooth along the curve. Consider the partial derivatives restricted to $C$ (that is, take $Q_{ij} = 0$ and $x_l = 0$ for $l\\ge e+1$)\n\\[ \\frac{\\partial F}{\\partial x_2} = -x_0 + 2\\lambda_2x_2 - \\lambda_3x_4, \\ \\ \\frac{\\partial F}{\\partial x_{n}} = x_e. \\]\nIf $P = (s^e : s^{e-1}t : \\cdots : t^e : 0 : \\cdots : 0)\\in C$ is a singular point of $X$, then $\\frac{\\partial F}{\\partial x_{n}} = x_e = t^e = 0$ implies $t=0$, thus $\\frac{\\partial F}{\\partial x_2} = -x_0 + 2\\lambda_2x_2 - \\lambda_3x_4 = -s^e = 0$, so $s=0$. Hence, $X$ is smooth along $C$.\n\n\\end{proofofthm4.3}\n\nA natural question that arises is whether we can obtain all splitting types from smooth quadric hypersurfaces $X$. The construction above produce quadrics smooth along $C$, but not necessarily smooth. We can, however, use the quadratic form matrix of $F$ to reduce the dimension of the singular locus of $X$.\n\n\n\\begin{theorem}\\label{corankexample}\nFor every splitting type\n\\[ E = \\left ( \\bigoplus_{i=1}^{e-2}\\mathcal{O}(e+2-a_i) \\right )\\oplus \\left ( \\bigoplus _{j=e+1}^{n}\\mathcal{O}(e-b_j)\\right ), \\]\nwith $a_i, b_j\\ge 0$ and $\\sum_{i=1}^{e-2}a_i + \\sum_{j=e+1}^nb_j = e-2$, we obtain an example of quadric $X$ of corank at most $\\sum _{a_i\\ge 4}(a_i-3)$ with $N_{C/X}\\cong E$. In particular, if $a_i\\le 3$ for all $i$, there exists a smooth quadric $X$ with $N_{C/X}\\cong E$.\n\\end{theorem}\n\\begin{proof} We first refer to the proof of Theorem \\ref{quadricprop1}, as we will use the quadric $F$ constructed there. We divide the proof into the cases $e=n$ and $e<n$.\n\nLet $e=n$. Rearrange the $a_i$ in increasing order $a_1\\le \\cdots \\le a_{n-2}$ and follow the construction of $F$ in the proof of Theorem \\ref{quadricprop1}. For $\\beta_0 = 0$ and $\\beta_i = a_1 + \\cdots + a_i$ for $i=1, \\cdots , n-2$, we obtain the polynomial\n\\[ F = \\sum_{i=0}^{n-2}Q_{\\beta_i+1,\\beta_i+2} =  Q_{1,2} + Q_{\\beta_1 + 1, \\beta_1 +2} + Q_{\\beta_2 + 1, \\beta_2 + 2} \\cdots + Q_{\\beta_{n-3}+1, \\beta_{n-3}+2} + Q_{n-1,n}, \\]\ncorresponding to the map \n\\[ \\psi_F = (s^{n-2}, \\ 0, \\dots , 0, \\ s^{n-2-\\beta_1}t^{\\beta_1},\\ 0, \\dots , 0,\\  s^{n-2-\\beta_2}t^{\\beta_2}, \\ \\ \\cdots \\ \\ , \\ s^{n-2-\\beta_{n-3}}t^{\\beta_{n-3}}, \\ 0, \\dots , 0 , \\ t^{n-2}), \\]\nwhose kernel is $N_{C/X}\\cong \\bigoplus_{i=1}^{n-2}\\mathcal{O}(n+2-a_i)$.\n\nWe look at the quadratic form matrix of $F$. It is the $(n+1)\\times (n+1)$ symmetric matrix $Q = (c_{i,j})_{i,j=0}^n$ with entries $c_{\\beta_{i}+1,\\beta_{i}+1} = 1$, $c_{\\beta_{i},\\beta_{i}+2} = -\\frac{1}{2}$, $c_{\\beta_{i}+2,\\beta_{i}} = -\\frac{1}{2}$, $0\\le i\\le n-2$, and zero elsewhere. It has 1's and 0's in the diagonal with $a_i-1$ consecutive 0's for each $a_i$, in increasing order of $i$. For example, for $n=5$ and $N_{C/X}\\cong \\mathcal{O}(5)\\oplus \\mathcal{O}(6)\\oplus \\mathcal{O}(7)$, we have $a_1=0$, $a_2=1$ and $a_3=2$, and we construct $\\psi_F = (s^3,\\ s^2t,\\ 0,\\ t^3)$, that corresponds to the matrix\n\\[ Q = \\left (\\begin{smallmatrix}\n0 & 0 & -\\frac{1}{2} & 0 & 0 & 0 \\\\\n0 & 1 & 0 & -\\frac{1}{2} & 0 & 0 \\\\\n-\\frac{1}{2} & 0 & 1 & 0 & 0 & 0 \\\\\n0 & -\\frac{1}{2} & 0 & 0 & 0 & -\\frac{1}{2} \\\\\n0 & 0 & 0 & 0 & 1 & 0 \\\\\n0 & 0 & 0 & -\\frac{1}{2} & 0 & 0 \n\\end{smallmatrix}\\right ). \\]\n\n\nNotice that, for each $a_i\\ge 4$, we have a $(a_i-3)\\times (a_i-3)$ zero diagonal block in $Q$. Thus, the rank of $F$ drops by at least $a_i-3$ for each relation of degree $a_i\\ge 4$. The nonzero diagonal blocks might be singular, further reducing the rank of $F$. To prove the theorem, we will show we can replace the nonzero blocks by nonsingular ones without changing the splitting type of $N_{C/X}$.\n\nFirst, notice that the first diagonal block of $Q$ corresponds to the $a_i\\le 2$. This block is followed by $(a_i-3)\\times (a_i-3)$ zero diagonal blocks alternating with blocks\n\\[ B = \\begin{pmatrix}\n0 & 0 & -\\frac{1}{2}\\\\\n0 & 1 & 0\\\\\n-\\frac{1}{2} & 0 & 0\n\\end{pmatrix}. \\]\n\nThe blocks B are already nonsingular, so we do not need to replace them. We might need to replace the first block. When we replace it by a new block, we change $F$ to an $F'$, and we need to check if the new $\\psi_{F'}$ induces the same bundle $N_{C/X}$. We do this by checking that $\\psi_{F'}$ has column relations of the same degree as of $\\psi_F$.\n\nLet us analyze the first block closer. It has size $(m+1)\\times (m+1)$, for $m = \\sum_{a_i\\le 2}a_i + 2$. Its diagonal is formed by 0, a sequence of consecutive 1's, and then alternating 1's and 0's:\n\\[ B_1 = \\left ( \\begin{smallmatrix}\n0 &  & -\\frac{1}{2} &  &  &  &  &  &  &  & \\\\\n & 1 &  & -\\frac{1}{2} &  &  &  &  &  &  & \\\\\n-\\frac{1}{2} &  & 1 &  &  &  &  &  &  &  & \\\\\n & -\\frac{1}{2} &  & ... &  & -\\frac{1}{2} &  &  &  &  & \\\\\n &  &  &  & 1 &  &  &  &  &  & \\\\\n &  &  & -\\frac{1}{2} &  & 0 &  & -\\frac{1}{2} &  &  & \\\\\n &  &  &  &  &  & 1 &  &  &  & \\\\\n &  &  &  &  & -\\frac{1}{2} &  & ... &  &  & \\\\\n &  &  &  &  &  &  &  & 0 &  & -\\frac{1}{2}\\\\\n &  &  &  &  &  &  &  &  & 1 & \\\\\n &  &  &  &  &  &  &  & -\\frac{1}{2} &  & 0\n\\end{smallmatrix}\\right ). \\]\n\nLet $l = |\\{ i \\ | \\ a_i = 1 \\}| + 1$ be the number of consecutive 1's in the diagonal. We divide into the cases $l=1, 2, 3$ and $l\\ge 4$;\n\nFor the case $l=1$, the diagonal of $B_1$ alternates between $0$ and $1$, thus $m$ must be even. If $4\\nmid m$, Gauss-Jordan elimination shows that the original matrix is nonsingular and we do not change it. If $4\\mid m$, it may be singular, and we define $F' = F + Q_{\\frac{m}{2}, \\frac{m}{2}+1} + Q_{\\frac{m}{2}-1, \\frac{m}{2}+1}$. This change will replace the three middle entries of $\\psi_F$:\n\\[ \\psi_F = (s^{n-2}, \\ 0, \\ s^{n-4}t^2, \\ 0, \\cdots , s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2}, \\ 0, \\ s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}}, \\cdots ,\\ 0, \\ s^{n-m}t^{m-2} \\cdots ) \\]\nby\n\\begin{multline*}\n    \\psi_{F'} = (s^{n-2}, \\ 0, \\ s^{n-4}t^2, \\ 0, \\cdots , s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2} + s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}}, \\ s^{n-\\frac{m}{2}-1}t^{\\frac{m}{2}-1}, \\ s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}} + s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2}, \\cdots \\\\\n    \\cdots ,\\ 0, \\ s^{n-m}t^{m-2} \\cdots ),\n\\end{multline*}\nand instead of the original column relations\n\\begin{align*}\n    & t^2(s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2}) - s^2(s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}}) = 0\\\\\n    & 1\\cdot (s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2}) - 1\\cdot 0 = 0,\n\\end{align*}\nwe get\n\\begin{align*}\n    & (st)(s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2} + s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}}) - (s^2+t^2)(s^{n-\\frac{m}{2}-1}t^{\\frac{m}{2}-1}) = 0\\\\\n    & 1\\cdot (s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2} + s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}}) - 1\\cdot (s^{n-\\frac{m}{2}-2}t^{\\frac{m}{2}} + s^{n-\\frac{m}{2}}t^{\\frac{m}{2}-2}) = 0,\n\\end{align*}\nthus preserving the degrees of the column relations.\n\n\n\n\n\n\nFor $l=2$ or $l\\ge 4$, Gauss-Jordan elimination shows that $B_1$ is nonsingular, and we do not replace it.\n\nFor $l=3$, consider the cases $m=4$ and $m>4$. If $m>4$, switch one of the 0's to the third row:\n\\[ \\text{replace } \\left ( \\begin{smallmatrix}\n0 &  & -\\frac{1}{2} &  &  &  &  & &\\\\\n & 1 &  & -\\frac{1}{2} &  &  &  & &\\\\\n-\\frac{1}{2} &  & 1 &  & -\\frac{1}{2} &  &  & &\\\\\n & -\\frac{1}{2} &  & 1 &  &  &  & &\\\\\n &  & -\\frac{1}{2} &  & 0 &  & -\\frac{1}{2} & &\\\\\n &  &  &  &  & 1 &  & &\\\\\n &  &  &  & -\\frac{1}{2} &  & 0 & & -\\frac{1}{2}\\\\\n &  &  &  &  &  &  & 1 &\\\\\n &  &  &  &  &  & -\\frac{1}{2} &  &...\n\\end{smallmatrix}\\right )\\text{ by } \n\\left ( \\begin{smallmatrix}\n0 &  & -\\frac{1}{2} &  &  &  &  & &\\\\\n & 1 &  &  &  &  &  & &\\\\\n-\\frac{1}{2} &  & 0 &  & -\\frac{1}{2} &  &  & &\\\\\n &  &  & 1 &  & -\\frac{1}{2} &  & &\\\\\n &  & -\\frac{1}{2} &  & 1 &  & -\\frac{1}{2} & &\\\\\n &  &  & -\\frac{1}{2} &  & 1 &  & &\\\\\n &  &  &  & -\\frac{1}{2} &  & 0 & & -\\frac{1}{2}\\\\\n &  &  &  &  &  &  & 1 &\\\\\n &  &  &  &  &  & -\\frac{1}{2} &  &...\n\\end{smallmatrix}\\right ).\\]\n\nThe second matrix is nonsingular, as one can also see by Gauss-Jordan elimination. In terms of the polynomial, this corresponds to taking $F' = F - Q_{2,3} + Q_{4,5}$, and it replaces\n\\[ \\psi_F = ( s^{n-2},\\ s^{n-3}t, \\ s^{n-4}t^2, \\ 0, \\ s^{n-6}t^4, \\ 0, \\ s^{n-8}t^6, \\cdots ) \\]\nby\n\\[ \\psi_F' = ( s^{n-2},\\ 0, \\ s^{n-4}t^2, \\ s^{n-5}t^3, \\ s^{n-6}t^4, \\ 0, \\ s^{n-8}t^6, \\cdots ). \\]\nNote that this only changes the position of one degree $2$ relation, that is, instead of the column relations\n\\begin{align*}\n    & t(s^{n-2}) - s(s^{n-3}t) = 0\\\\\n    & t(s^{n-3}t) - s(s^{n-4}t^2) = 0\\\\\n    & t^2(s^{n-4}t^2) - s^2(s^{n-6}t^4) = 0,\n\\end{align*}\nwe get the relations\n\\begin{align*}\n    & t^2(s^{n-2}) - s^2(s^{n-4}t^2) = 0\\\\\n    & t(s^{n-4}t^2) - s(s^{n-5}t^3) = 0\\\\\n    & t(s^{n-5}t^3) - s(s^{n-6}t^4) = 0.\n\\end{align*}\n\nIf $m=4$, the first block of the original matrix $Q$ is\n\\[ \\left (\\begin{smallmatrix}\n0 &  & -\\frac{1}{2} &  & \\\\\n & 1 &  & -\\frac{1}{2} & \\\\\n-\\frac{1}{2} &  & 1 &  & -\\frac{1}{2}\\\\\n & -\\frac{1}{2} &  & 1 & \\\\\n &  & -\\frac{1}{2} &  & 0\n\\end{smallmatrix}\\right ). \\]\n\nWe replace it by the nonsingular\n\\[ \\left (\\begin{smallmatrix}\n0 &  & -\\frac{1}{2} & -\\frac{1}{2} & -\\frac{1}{2}\\\\\n & 1 & \\frac{1}{2} & 0 & \\\\\n-\\frac{1}{2} & \\frac{1}{2} & 1 &  & -\\frac{1}{2}\\\\\n-\\frac{1}{2} & 0 &  & 1 & \\\\\n-\\frac{1}{2} &  & -\\frac{1}{2} &  & 0\n\\end{smallmatrix}\\right ), \\]\ncorresponding to the polynomial $F' = F + Q_{1,3} + Q_{1,4}$. The original map $\\psi_F$ is\n\\[ \\psi_F = ( s^{n-2},\\ s^{n-3}t, \\ s^{n-4}t^4, \\ 0, \\cdots ) \\]\nand satisfies the degree $1$ column relations\n\\begin{align*}\n    & t(s^{n-2}) - s(s^{n-3}t) = 0 \\\\\n    & t(s^{n-3}t) - s(s^{n-4}t^2) = 0.\n\\end{align*}\nThe new map is\n\\[ \\psi_{F'} = (s^{n-2} + s^{n-3}t + s^{n-4}t^2, \\ s^{n-2}+2s^{n-3}t, \\ s^{n-2} + s^{n-4}t^2, \\cdots) \\]\nand satisfies the degree $1$ relations\n\\begin{align*}\n    & (-s+3t)(s^{n-2} + s^{n-3}t + s^{n-4}t^2) + (s-t)(s^{n-2}+2s^{n-3}t) - (3t)(s^{n-2} + s^{n-4}t^2)) = 0 \\\\\n    & (-s-2t)(s^{n-2} + s^{n-3}t + s^{n-4}t^2) + t(s^{n-2}+2s^{n-3}t) + (s+2t)(s^{n-2} + s^{n-4}t^2) = 0,\n\\end{align*}\nconfirming that replacing $F$ by $F'$ does not change the degrees of the column relations.\n\nSo far, we have shown that we can replace the first block of the matrix of $F$ by a nonsingular block with column relations in $\\psi_F$ of the same degrees. We now need to check the column relation between the first and the other blocks. It is enough to check the relation between the first and the second nonzero block.\n\nExcept for the case $l = 3$ and $m = 4$, the last entry of $\\psi_F$ corresponding to the first block remains the same, and so do the column relation between the first and the second blocks. When $l = 3$ and $m = 4$, the map $\\psi_F$ looks like\n\\[ \\psi_{F} = (s^{n-2}, \\ s^{n-3}t, \\ s^{n-4}t^2, \\ 0, \\cdots , 0, \\ s^{n-2-b}t^b, \\cdots), \\]\nand the relation between the first and second blocks is the degree $b-2$ relation\n\\[ t^{b-2}(s^{n-4}t^2) - s^{b-2}(s^{n-2-b}t^b) = 0. \\]\n\nIt gets replaced by\n\\[ \\psi_{F'} = (s^{n-2} + s^{n-3}t + s^{n-4}t^2, \\ s^{n-2}+2s^{n-3}t, \\ s^{n-2} + s^{n-4}t^2, \\ 0, \\cdots , 0, \\ s^{n-2-b}t^b, \\cdots), \\]\nand we have the relation between the first and the second blocks\n\\[ t^{b-2}(2(s^{n-2} + s^{n-3}t + s^{n-4}t^2) - (s^{n-2}+2s^{n-3}t) - (s^{n-2} + s^{n-4}t^2)) - s^{b-2}(s^{n-2-b}t^b) = 0, \\]\nalso of degree $b-2$.\n\nTherefore, the new matrix we obtain by replacing the first block induces column relations of the same degree of the original matrix. Hence, we still get $N_{C/X}\\cong \\bigoplus _{i=1}^{n-2}\\mathcal{O}(n+2-a_i)$. The rank of the new matrix is only decreased by the $(a_i-3)\\times (a_i-3)$ zero diagonal blocks, thus it has corank $\\sum _{a_i\\ge 4}(a_i-3)$.\n\n\\medskip\n\nNow, let $e<n$. The $F$ constructed in Proposition \\ref{quadricprop1} is $F = \\sum_{i=0}^{e-2}Q_{\\beta_i+1,\\beta_i+2} + \\sum_{j=e+1}^n x_{\\gamma_j}x_j$. The quadratic form matrix of $F$ can be seen as a four blocks symmetric matrix\n\\[ M = \\begin{pmatrix}\nQ & A \\\\\nA^t & 0\n\\end{pmatrix} \\]\nwhere $Q$ is the $(e+1)\\times (e+1)$ matrix corresponding to the quadratric form $\\sum_{i=0}^{e-2}Q_{\\beta_i+1,\\beta_i+2}$ in $\\P^e = V(x_{e+1}, \\cdots , x_n)$. Thus, by applying the case $e=n$ to the matrix $Q$, we may assume $Q$ to be a corank $\\sum _{a_i\\ge 4}(a_i-3)$ matrix. In other words, we are able to replace $Q$ by a corank $\\sum _{a_i\\ge 4}(a_i-3)$ matrix with the same normal bundle $N_{C/X}$.\n\nThe $0$ block of $M$ corresponds to the terms $x_lx_j$ of $F$ with $l,j\\ge e+1$. These are $0$ when restricted to $C$, and therefore do not interfere with the map $\\psi_F$. Thus, we can replace the $0$ block by any matrix $L$ without changing $\\psi_F$ and hence preserving $N_{C/X}$. Therefore, all matrices\n\\[ M = \\begin{pmatrix}\nQ & A \\\\\nA^t & L\n\\end{pmatrix} \\]\nwith any $L$ induce the same $N_{C/X}$. To compute its rank, assume $L$ is invertible, let $I$ be the identity matrix, and use Schur complement:\n\\[ \\begin{pmatrix}\nI & -AL^{-1} \\\\\n0 & I\n\\end{pmatrix}\n\\begin{pmatrix}\nQ & A \\\\\nA^t & L\n\\end{pmatrix}\n\\begin{pmatrix}\nI & 0\\\\\n-L^{-1}A^t & I\n\\end{pmatrix} = \n\\begin{pmatrix}\nQ-AL^{-1}A^t & 0 \\\\\n0 & L\n\\end{pmatrix}. \\]\n\nSince \n\\[ \\begin{pmatrix}\nI & -AL^{-1} \\\\\n0 & I\n\\end{pmatrix} \\ \\ \\text{ and } \\ \\ \\begin{pmatrix}\nI & 0\\\\\n-L^{-1}A^t & I\n\\end{pmatrix} \\]\nare invertible, it follows that\n\\[ \\rk \\begin{pmatrix}\nQ & A \\\\\nA^t & L\n\\end{pmatrix} = \\rk \\begin{pmatrix}\nQ-AL^{-1}A^t & 0 \\\\\n0 & L\n\\end{pmatrix} = \\rk (Q-AL^{-1}A^t) + \\rk (L). \\]\n\nAnd since matrices of rank at least $\\rk (Q)$ form an open neighborhood of $Q$, we can choose an invertible matrix $L$ such that $\\rk (Q-AL^{-1}A^t)\\ge \\rk (Q)$, and therefore, $\\rk M\\ge \\rk(Q) + \\rk(L)$, that is, the corank of $M$ is at most $\\sum _{a_i\\ge 4}(a_i-3)$.\n\n\\end{proof}\n\n\\printbibliography[title={References}]\n\n\\end{document}\n\n", "meta": {"timestamp": "2022-09-28T02:12:50", "yymm": "2209", "arxiv_id": "2209.13199", "language": "en", "url": "https://arxiv.org/abs/2209.13199", "abstract": "Let $C$ be the rational normal curve of degree $e$ in $\\mathbb{P}^n$, and let $X\\subset \\mathbb{P}^n$ be a degree $d\\ge 2$ hypersurface containing $C$. In previous work, I. Coskun and E. Riedl showed that the normal bundle $N_{C/X}$ is balanced for a general $X$. H. Larson studied the case of lines ($e=1$) and computed the dimension of the space of hypersurfaces for which $N_{C/X}$ has a given splitting type. In this paper, we work with any $e\\ge 2$. We compute explicit examples of hypersurfaces for all possible splitting types, and for $d\\ge 3$, we compute the dimension of the space of hypersurfaces for which $N_{C/X}$ has a given splitting type. For $d=2$, we give a lower bound on the maximum rank of quadrics with fixed splitting type.", "subjects": "Algebraic Geometry (math.AG)", "title": "Normal Bundles of Rational Normal Curves on Hypersurfaces", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9865717468373085, "lm_q2_score": 0.8539127455162773, "lm_q1q2_score": 0.8424461889906357}}
{"url": "https://arxiv.org/abs/1401.7156", "title": "On weighted Hardy inequalities for non-increasing sequences", "abstract": "A result of Bennett and Grosse-Erdmann characterizes the weights for which the corresponding weighted Hardy inequality holds on the cone of non-negative, non-increasing sequences and a bound for the best constant is given. In this paper, we improve the bound for $1<p \\leq 2$.", "text": "\\section{Introduction}\n\\label{sec 1} \\setcounter{equation}{0}\n\n  Throughout this paper, we let $p \\geq 1$. For $p \\neq 1$  we let $q$ be defined by $\\frac{1}{p}+\\frac{1}{q}=1$ and we set $1/q=0$ when $p=1$. Consider the following weighted Hardy inequality on the cone of non-negative, non-increasing sequences ${\\bf x}=(x_n)_{n \\geq 1}$:\n\\begin{align}\n\\label{1.01}\n \\sum^{\\infty}_{n=1}b_n \\left (\\sum^{n}_{k=1}\\frac {x_k}{n} \\right\n   )^p \\leq U_p\\sum^{\\infty}_{n=1}b_nx^p_n,\n\\end{align}\n   where $(b_n)_{n \\geq 1}$ is a non-negative sequence, $U_p>0$ a constant independent of ${\\bf x}$. In \\cite[Theorem 1]{BGE}, Bennett and Grosse-Erdmann gave a complete characterization on the sequence $(b_n)_{n \\geq 1}$ such that $U_p$ exists. They showed that this is the case if and only if there exists a constant $U'_p>0$ such that for all $n \\geq 1$,\n\\begin{equation*}\n   \\sum^{\\infty}_{k=n}\\frac {b_k}{k^p} \\leq \\frac {U'_p}{n^p}\\sum^{n}_{k=1}b_k.\n\\end{equation*}\n   Moreover, if the constants $U_p, U'_p$ are chosen best possible, then\n\\begin{align}\n\\label{1.2}\n   U'_p \\leq U_p \\leq p^p(U'_p+1)^p.\n\\end{align}\n\n   Integral inequalities analogous to \\eqref{1.01} for non-increasing functions have been studied by Ari\\~no and Muckenhoupt in \\cite{AM}. They showed that if $p \\geq 1$ and $v$ is a non-negative measurable\nfunction on $(0, \\infty)$ then there is a constant $V_p>0$ such that\n\\begin{align*}\n  \\int^{\\infty}_0v(x)\\left ( \\frac 1x\\int^x_0f(t)dt\\right )^p dx \\leq V_p\\int^{\\infty}_{0}v(x)f^p(x)dx\n\\end{align*}\n   holds for all non-negative non-increasing functions $f(x)$ if and only if there is a constant $V'_p>0$ such that for all $x>0$,\n\\begin{align*}\n  \\int^{\\infty}_x\\frac{v(t)}{t^p}dt \\leq \\frac {V'_p}{x^p}\\int^{x}_{0}v(t)dt.\n\\end{align*}\n   The argument of Bennett and Grosse-Erdmann also works for the integral case and it implies that \\cite[(17)]{BGE} if the constants $V_p, V'_p$ are chosen best possible, then\n\\begin{align*}\n   V'_p \\leq V_p \\leq (V'_p+1)^p.\n\\end{align*}\n\n   Comparing the above two results, we see that in the discrete case, the corresponding bounds for the best constants are not as good as what is given in the integral case. It is then natural to seek for an improvement on the bounds given in \\eqref{1.2}, which is the goal of this paper.\n   Our result in this paper is the following generalization of the above mentioned result of Bennett and Grosse-Erdmann:\n\\begin{theorem}\n\\label{thm1}\n   Let $p \\geq 1$ be fixed. Let $(b_n)_{n \\geq 1}$ be a non-negative sequence and let $(\\lambda_n)_{n \\geq 1}$ be a non-negative,\n    non-increasing sequence with $\\lambda_1 >0$. Let $\\Lambda_n=\\sum^n_{k=1}\\lambda_k$.\n   Then there is a constant $U_p > 0$ such that\n\\begin{align}\n\\label{1.0}\n   \\sum^{\\infty}_{n=1}b_n \\left (\\sum^{n}_{k=1}\\frac {\\lambda_kx_k}{\\Lambda_n} \\right\n   )^p \\leq U_p\\sum^{\\infty}_{n=1}b_nx^p_n\n\\end{align}\n  holds for all non-negative, non-increasing sequences $(x_n)_{n \\geq 1}$ if and\nonly if there is a constant $U'_p > 0$ such that for all $n \\geq 1$,\n\\begin{align}\n\\label{1.0'}\n   \\sum^{\\infty}_{k=n}\\frac{b_k}{\\Lambda^p_k} \\leq \\frac\n   {U'_p}{\\Lambda^p_n}\\sum^{n}_{k=1}b_k.\n\\end{align}\n   Moreover, if $U_p$ and $U'_p$ are chosen best-possible then we have\n\\begin{align}\n\\label{1.7}\n   U'_p \\leq U_p \\leq  \\left\\{\\begin{array}{ll}\n\\left (pU'_p+1 \\right )^p, & 1 \\leq p \\leq 2;   \\\\\n p^p(U'_p+1)^p, & p >2.\n\\end{array}\\right.\n\\end{align}\n\\end{theorem}\n\n   The case $\\lambda_n=1$ of Theorem \\ref{thm1} gives back the result of\n   Bennett and Grosse-Erdmann except that instead of \\eqref{1.7}, the upper\n   bound given for $U_p$ in \\cite[Theorem 1]{BGE} is given as in \\eqref{1.2} for all $p \\geq 1$. Theorem \\ref{thm1} therefore improves upon the result of Bennett and Grosse-Erdmann for $1 < p \\leq 2$ in this sense.\nWe point out here that this improvement comes from our refinement (see Lemma \\ref{lem1}) on the so called ``Power Rule\" (Lemma \\ref{lem0} below), a key lemma used in the proof of \\cite[Theorem 1]{BGE} by Bennett and Grosse-Erdmann.\n\n  \n\\section{lemmas}\n\\label{sec 2} \\setcounter{equation}{0}\n   \n\\begin{lemma}[{\\cite[Lemma 3]{BGE}}]\n\\label{lem0} Let $p \\geq 1$. Then for all non-negative sequences $(a_k)_{k \\geq 1}$, any integer $n \\geq 1$,\n\\begin{align*}\n   \\left ( \\sum^{\\infty}_{k=n}a_k\\right )^p \\leq p\\sum^{\\infty}_{k=n}a_k\\left ( \\sum^{\\infty}_{i=k} a_i\\right\n   )^{p-1}.\n\\end{align*}\n\\end{lemma}\n\n\\begin{lemma}[{{\\cite[Lemma 2]{BGE}}}]\n\\label{lem2}\n    Let $(u_n)_{n \\geq 1}, (v_n)_{n \\geq 1}$ be two non-negative sequences satisfying for any integer $n \\geq 1$,\n\\begin{equation*}\n   \\sum_{i=1}^n u_i \\leq \\sum_{i=1}^n v_i,\n\\end{equation*}\n    then for all non-negative, non-increasing sequences $(a_n)_{n \\geq 1}$,\n\\begin{equation*}\n   \\sum_{i=1}^n u_i a_i \\leq \\sum_{i=1}^n v_i a_i.\n\\end{equation*}\n\\end{lemma}\n\n\\begin{lemma}[{\\cite[Lemma 3.1]{G7}}]\n\\label{lem3}\n   Let $(B_n )_{n\\geq 1}$ and $( C_n )_{n \\geq 1}$ be strictly increasing positive sequences with\n   $B_1/B_2 \\leq C_1 / C_2$. If for any integer $n \\geq 1$,\n\\begin{equation*}\n  \\frac {B_{n+1}-B_n}{B_{n+2}-B_{n+1}} \\leq  \\frac\n  {C_{n+1}-C_n}{C_{n+2}-C_{n+1}}.\n\\end{equation*}\n  Then $B_{n}/B_{n+1} \\leq C_{n} / C_{n+1}$ for any integer $n \\geq 1$.\n\\end{lemma}\n\\begin{lemma}\n\\label{lem2.2} Let $1 \\leq p \\leq 2$ and let $n \\geq 1$ be a fixed integer. Let $\\lambda=(\\lambda_k)_{1 \\leq k \\leq n}$ be a non-negative, non-increasing sequence with $\\lambda_1>0$. For $1 \\leq k \\leq n$, let $\\Lambda_k=\\sum^k_{i=1}\\lambda_i$ and\n\\begin{equation*}\n    C_{k,p,\\lambda}=\\frac\n   {\\Lambda^p_k}{\\sum^{k}_{i=1}\\lambda_i\\Lambda^{p-1}_i}.\n\\end{equation*}\nThen the sequence $(C_{k,p,\\lambda})_{1 \\leq k \\leq n}$ is increasing with respect\nto $k$.\n\\end{lemma}\n\\begin{proof}\n   The assertion holds trivially when $p=1$, so we may assume $p>1$. We may assume $n \\geq 2$ and $\\lambda_k>0$ for all $1 \\leq k \\leq n$. We extend the sequence $\\lambda$ to be indexed by all positive integers by defining $\\lambda_i=\\lambda_n/i$ for $i \\geq n+1$. We define similarly $\\Lambda_k, C_{k,p,\\lambda}$ for $k > n$. It therefore suffices to show that $C_{k, p, \\lambda} \\leq C_{k+1, p, \\lambda}$\n   for all $k \\geq 1$. Applying Lemma \\ref{lem3} with $B_k=\\Lambda^p_k, C_k=\\sum^{k}_{i=1}\\lambda_i\\Lambda^{p-1}_i$, \n   ones checks directly that $B_1/B_2 \\leq C_1/C_2$. Thus, it remains to\n   show for that all $k \\geq 1$,\n\\begin{align*}\n   \\frac {\\Lambda^p_{k+1}-\\Lambda^p_{k}}{\\lambda_{k+1}\\Lambda^{p-1}_{k+1}}\n   \\leq \\frac\n   {\\Lambda^p_{k+2}-\\Lambda^p_{k+1}}{\\lambda_{k+2}\\Lambda^{p-1}_{k+2}}.\n\\end{align*}\n   When we regard $\\lambda_{k+2}$ as a variable with $0 \\leq\n   \\lambda_{k+2} \\leq \\lambda_{k+1}$, then it is easy to see that\n   the right-hand side expression above is a decreasing function of\n   $\\lambda_{k+2}$ and hence it suffices to show that the above\n   inequality holds with $\\lambda_{k+2}=\\lambda_{k+1}$. In this\n   case, on setting $\\lambda_{k+1}=x, \\Lambda_k=y$ with $y \\geq\n   x$, we can recast the above inequality as\n\\begin{align*}\n   x-(x+y)^p(2x+y)^{1-p}+y^p(x+y)^{1-p} \\geq 0.\n\\end{align*}\n   We further set $z=x/y$ to recast the above inequality as\n\\begin{align*}\n   z-(1+z)^p(1+2z)^{1-p}+(1+z)^{1-p} \\geq 0.\n\\end{align*}\n   Upon dividing $1+z$ on both sides of the above inequality and\n   setting $t=z/(1+z)$, we see that it suffices to show for $0\n   \\leq t \\leq 1/2$,\n\\begin{align*}\n   g(t):=t-(1+t)^{1-p}+(1-t)^p \\geq 0.\n\\end{align*}\n   It's easy to see that $g(0)=g'(0)=0$ and\n   $g''(t)=p(p-1)((1-t)^{p-2}-(1+t)^{-p-1}) \\geq 0$ when $1<p \\leq 2$. This implies that $g(t)$ is an increasing function of $0 \\leq t \\leq 1/2$ which \n   completes the proof.\n\\end{proof}\n\\begin{lemma}\n\\label{lem1} Let $p \\geq 1$, $\\lambda=(\\lambda_k)_{k \\geq 1}$ a\nnon-negative, non-increasing sequence with $\\lambda_1>0$.  Then for all non-negative,\nnon-increasing sequences $(a_k)_{k \\geq 1}$, any integer $n \\geq 1$,\n\\begin{equation}\n\\label{2.1}\n   \\left ( \\sum^n_{k=1}\\lambda_ka_k\\right )^p \\leq C_{n,p, \\lambda}\\sum^n_{k=1}\\lambda_ka_k\\left ( \\sum^k_{i=1}\\lambda_ia_i\\right\n   )^{p-1},\n\\end{equation}\n   where $C_{n, p,\\lambda}$ is defined as in Lemma \\ref{lem2.2} when $1 \\leq p \\leq 2$ and $C_{n,p,\\lambda}=p$ when $p>2$. Moreover, when $1 \\leq p \\leq 2$, the constant $C_{n,p,\\lambda}$ is best possible and equality in\n    \\eqref{2.1} holds when $1<p \\leq 2$ if and only if $a_1=a_2=\\ldots=a_n$.\n\\end{lemma}\n\\begin{proof}\n    As inequality \\eqref{2.1} follows from Lemma \\ref{lem0} when $p>2$ and the assertion of the lemma follows trivially for $p=1$, we only need to consider the case $1<p \\leq 2$. We define\n\\begin{equation*}\n    f_n(x_1, x_2, \\ldots, x_n)=\\left ( \\sum^n_{k=1}\\lambda_kx_k\\right )^p -C_{n,p,\\lambda}\\sum^n_{k=1}\\lambda_kx_k\\left ( \\sum^k_{i=1}\\lambda_ix_i\\right\n   )^{p-1}.\n\\end{equation*}\n    By homogeneity, it suffices to show $f_n \\leq 0$ on the compact set $\\{ (x_1, \\ldots, x_n) | 1 \\geq\n    x_1 \\geq x_2 \\geq \\ldots \\geq x_n \\geq 0 \\}$. We may assume $\\lambda_k>0$ for all\n    $k$ here as discarding the zero terms and relabeling will not change the expression.\n\n   As $f_1=0$ holds trivially, we may assume $n \\geq 2$ here.  Assume the maximum of $f_n$ is\n    attained at some ${\\bf x}_0=\\left ( \\left( {\\bf x}_0 \\right)_1, \\left( {\\bf x}_0 \\right)_2, \\ldots, \\left( {\\bf x}_0 \\right)_n\n    \\right)$ with $\\left( {\\bf x}_0 \\right)_1 \\geq \\left( {\\bf x}_0 \\right)_2 \\geq\n\\ldots \\geq \\left( {\\bf x}_0\n    \\right)_n$. If $\\left( {\\bf x}_0\n    \\right)_{m+1}=0$ for some $1 \\leq m <n$, then as $C_{m,p, \\lambda} \\leq C_{n,p, \\lambda}$ by Lemma \\ref{lem2.2}, it is easy to\n    see that we are reduced to the consideration of $f_m \\leq 0$.\n    Thus, we may further assume $\\left( {\\bf x}_0\n    \\right)_n >0$ here.\n\n     Suppose $\\left( {\\bf x}_0\n    \\right)_m >\\left( {\\bf x}_0\n    \\right)_{m+1}>0$ for some $1 \\leq m <n$.\n    In this case we must have $\\partial f_n/\\partial x_m({\\bf x}_0) \\geq 0$ since $\\partial f_n/\\partial x_m({\\bf x}_0) < 0$ means\n    decreasing the value of $\\left( {\\bf x}_0\n    \\right)_m$ will increase the value of\n    $f_n$, a contradiction. Similar argument implies that $\\partial f_n/\\partial x_{m+1}({\\bf x}_0) \\leq 0$.\n    Therefore, we conclude that we have\n\\begin{align*}\n    0 & \\leq \\frac {1}{\\lambda_m}\\frac {\\partial f_n}{\\partial x_m}({\\bf x}_0)-\\frac {1}{\\lambda_{m+1}}\\frac {\\partial f_n}{\\partial x_{m+1}}({\\bf\n    x}_0)  \\\\\n    &=C_{n,p, \\lambda}\\left ( \\left ( \\sum^{m+1}_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-1}-\\left ( \\sum^m_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-1}-(p-1)\\lambda_m\\left( {\\bf x}_0\n    \\right)_m\\left ( \\sum^m_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-2} \\right ).\n\\end{align*}\n   If $p=2$, this would imply $\\lambda_{m+1}\\left( {\\bf x}_0\n    \\right)_{m+1} \\geq \\lambda_m\\left( {\\bf x}_0\n    \\right)_m$, a contradiction. If $1<p<2$, by the Mean Value Theorem, we have\n\\begin{align*}\n    & \\left ( \\sum^{m+1}_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-1}-\\left ( \\sum^m_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-1} \\\\\n   =&(p-1)\\lambda_{m+1}\\left( {\\bf x}_0\n    \\right)_{m+1} \\xi^{p-2} <(p-1)\\lambda_{m}\\left( {\\bf x}_0\n    \\right)_m\\left ( \\sum^m_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i\\right\n   )^{p-2},\n\\end{align*}\n   as $\\sum^m_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i < \\xi < \\sum^{m+1}_{i=1}\\lambda_i\\left( {\\bf x}_0\n    \\right)_i$. This again leads to a contradiction. Thus we must have\n   $\\left( {\\bf x}_0 \\right)_1 =\\left( {\\bf x}_0\n\\right)_2=\\ldots=\\left( {\\bf x}_0 \\right)_n$, which implies that\n$f_n({\\bf x}_0)=0$ and the assertion\n   of the lemma follows for $1<p \\leq 2$.\n\\end{proof}\n\n    In what follows we make two remarks about Lemma \\ref{lem1}.\n    Throughout our remarks, we let $1 \\leq p \\leq 2$, $\\lambda_k=1$ for all $k$ with the function $f_n$ being defined as in the proof of Lemma\n    \\ref{lem1} and $C_{n,p,\\lambda}$ being defined as in\n    Lemma \\ref{lem2.2}.\n\n    Remark 1. For any given ${\\bf x}=(x_1,\n    x_2, \\ldots, x_n)$, we let ${\\bf x}'=(x_1, x_2, \\ldots, x_{i+1}, x_i, \\ldots, x_n)$ by permuting\n    two adjacent coordinates $x_i, x_{i+1}$ of ${\\bf x}$ for some $1 \\leq i\n    <n$, then we have\n\\begin{align*}\n   &f_n({\\bf x})-f_n({\\bf x}')  \\\\\n   =&-C_{n,p, \\lambda}\\left ( x_i(a+x_i)^{p-1}+x_{i+1}(a+x_i+x_{i+1})^{p-1}-x_{i+1}(a+x_{i+1})^{p-1}-x_{i}(a+x_i+x_{i+1})^{p-1} \\right\n   ),\n\\end{align*}\n   where we set (with empty sum being $0$)\n   $a=\\sum^{i-1}_{k=1}x_k$.\n\n    It is easy to check that the function\n    $S_r(x,y)=(x^r-y^r)/(x-y)$\n     is an increasing (respectively, decreasing) function of $y>0$ for fixed $0<x \\neq y$ when $r \\geq 1$ (respectively, $0<r \\leq 1$). Apply this with $r=p-1$, $x=a+x_i+x_{i+1}, y=a+x_i,\n     y'=a+x_{i+1}$, we see immediately that $f_n({\\bf x}) \\geq f_n({\\bf\n     x}')$ when $x_{i+1} \\geq x_i \\geq 0$ and $1<p \\leq 2$ or when $x_{i} \\geq  x_{i+1} \\geq 0$ and $p \\geq 2$.\n\n     It follows that when $p=2$ and $\\lambda_k=1$ for all $k$, the\n   maximum of $f_n$ on all non-negative sequences is the same as the maximum of $f_n$ on all non-negative,\nnon-increasing sequences. Thus, when $p=2,\\lambda_k=1$ for all $k$, the assertion of Lemma\n\\ref{lem1} holds for all non-negative sequences.\n\n     Remark 2.  We also remark that when $p>2$, we have for $n \\geq 2$,\n\\begin{align}\n\\label{2.2'}\n     \\frac {\\partial f_n}{\\partial x_n}\\left(\\left(1, 1, \\ldots, 1 \\right)\n     \\right)=n^{p-2}\\left ( np-C_{n,p,\\lambda}\\left(n+p-1 \\right ) \\right\n     ) <0,\n\\end{align}\n    where the last inequality is\n    equivalent to\n\\begin{equation*}\n    \\sum^{n}_{k=1}k^{p-1} <  \\frac {n^{p-1}}{p}(n+p-1), \\ n \\geq 2,\n\\end{equation*}\n     which in turn can be easily established by induction.\n\n     Inequality \\eqref{2.2'} implies that in this case\n    $0=f_n\\left(\\left(1, 1, \\ldots, 1 \\right)\n     \\right)<f_n\\left(\\left(1, 1, \\ldots, 1-\\epsilon \\right)\n     \\right)$ for some $\\epsilon>0$ small enough and this shows\n     that inequality \\eqref{2.1} does not hold for\n     all non-negative, non-increasing sequences when $p>2$.\n\n\n\\section{Proof of Theorem \\ref{thm1}}\n\\label{sec 3} \\setcounter{equation}{0}\n    We now proceed to the proof of Theorem \\ref{thm1}. Our\n    approach here follows that of Bennett and Grosse-Erdmann in their proof of \\cite[Theorem\n1]{BGE}. By considering the sequences $(1, \\ldots , 1,$$ 0, 0, $$\n\\ldots )$,  we see first that \\eqref{1.0'} is a necessary condition for\nthe validity of inequality \\eqref{1.0} and that $U'_p \\leq U_p$.\nConversely, assume that condition \\eqref{1.0'} holds. Note first that it follows from Lemma \\ref{lem0} and \\ref{lem1} that $C_{n,p,\\lambda} \\leq p$ where $C_{n,p,\\lambda}$ is defined as in Lemma \\ref{lem1}. Further\nnote that for any integer $n \\geq 1$,\n\\begin{align}\n\\label{2.3}\n   \\sum^n_{k=1} \\lambda_k\\Lambda^{p-1}_k\\sum^{\\infty}_{i=k}C_{i,p, \\lambda}\\frac {b_i}{\\Lambda^p_i}\n   & \\leq \\sum^n_{k=1} \\lambda_k\\Lambda^{p-1}_k\\sum^{n}_{i=k}C_{i,p,\n\\lambda}\\frac {b_i}{\\Lambda^p_i}\n   +\\sum^n_{k=1} \\lambda_k\\Lambda^{p-1}_k\\sum^{\\infty}_{i=n}C_{i,p, \\lambda}\\frac {b_i}{\\Lambda^p_i} \\\\\n  & \\leq \\sum^{n}_{i=1}C_{i,p, \\lambda}\\frac {b_i}{\\Lambda^p_i}\\sum^i_{k=1}\n  \\lambda_k\\Lambda^{p-1}_k+p\\sum^n_{k=1} \\lambda_k\\Lambda^{p-1}_k\\sum^{\\infty}_{i=n}\\frac\n  {b_i}{\\Lambda^p_i} \\nonumber \\\\\n  & \\leq \\sum^{n}_{i=1}C_{i,p, \\lambda}\\frac {b_i}{\\Lambda^p_i}\\sum^i_{k=1}\n  \\lambda_k\\Lambda^{p-1}_k+pU'_p\\frac {1}{\\Lambda^p_n}\\sum^n_{k=1}\n  \\lambda_k\\Lambda^{p-1}_k\\sum^{n}_{i=1}b_i \\nonumber \\\\\n  & \\leq  U''_p\\sum^n_{i=1}b_i, \\nonumber\n\\end{align}\n   where $U''_p=pU'_p+1, 1 \\leq p \\leq 2, U''_p=pU'_p+p, p>2$ and we have used \\eqref{1.0'} in the third inequality above and the bound $\\sum^n_{k=1}\n  \\lambda_k\\Lambda^{p-1}_k \\leq \\sum^n_{k=1}\n  \\lambda_k\\Lambda^{p-1}_n=\\Lambda^p_n$ in the last inequality above.\n\n Now by Lemma \\ref{lem1}, we have, for any non-negative, non-increasing sequences $(x_n)_{n \\geq 1}$,\n\\begin{align*}\n   \\sum^{\\infty}_{n=1}b_n \\left (\\sum^{n}_{k=1}\\frac {\\lambda_kx_k}{\\Lambda_n} \\right\n   )^p & \\leq \\sum^{\\infty}_{n=1}C_{n,p, \\lambda}\\frac {b_n}{\\Lambda^p_n}\\sum^{n}_{k=1}\\lambda_kx_k\\left (\\sum^{k}_{i=1}\\lambda_ix_i \\right\n   )^{p-1}  \\\\\n   &=\\sum^{\\infty}_{k=1}\\lambda_kx_k\\left(  \\sum^{\\infty}_{n=k}C_{n,p, \\lambda}\\frac {b_n}{\\Lambda^p_n} \\right )\\left (\\sum^{k}_{i=1}\\lambda_ix_i \\right\n   )^{p-1} \\\\\n   &= \\sum^{\\infty}_{k=1}\\left( \\lambda_k\\Lambda^{p-1}_k\\sum^{\\infty}_{n=k}C_{n,p, \\lambda}\\frac {b_n}{\\Lambda^p_n} \\right )x_k\n   \\left (\\sum^{k}_{i=1}\\frac {\\lambda_ix_i}{\\Lambda_k} \\right\n   )^{p-1}  \\\\\n   & \\leq U''_p\\sum^{\\infty}_{k=1}b_kx_k\n   \\left (\\sum^{k}_{i=1}\\frac {\\lambda_ix_i}{\\Lambda_k} \\right\n   )^{p-1}  \\\\\n   &= U''_p\\sum^{\\infty}_{k=1}b^{1/p}_kx_kb^{1/q}_k\n   \\left (\\sum^{k}_{i=1}\\frac {\\lambda_ix_i}{\\Lambda_k} \\right\n   )^{p-1},\n\\end{align*}\n    where the second inequality above follows from Lemma \\ref{lem2} and \\eqref{2.3}, the sequence\n    $$\\left(x_k\n   \\left (\\sum^{k}_{i=1}\\frac {\\lambda_ix_i}{\\Lambda_k} \\right\n   )^{p-1} \\right)_{k \\geq 1}$$ being non-negative, non-increasing.\n\n     By H\\\"older's inequality, we then have\n\\begin{align*}\n   \\sum^{\\infty}_{n=1}b_n \\left (\\sum^{n}_{k=1}\\frac {\\lambda_kx_k}{\\Lambda_n} \\right\n   )^p & \\leq  U''_p\\left (\\sum^{\\infty}_{n=1}b_nx^p_n \\right )^p\\left\n  (\\sum^{\\infty}_{k=1}b_k\n   \\left (\\sum^{k}_{i=1}\\frac {\\lambda_ix_i}{\\Lambda_k} \\right\n   )^p\\right )^{1/q},\n\\end{align*}\n    which implies \\eqref{1.0} with $U_p$ being replaced by ${U''_p}^p$ and this completes the\n  proof of Theorem \\ref{thm1}.\n\n\n\n", "meta": {"timestamp": "2014-01-29T02:09:05", "yymm": "1401", "arxiv_id": "1401.7156", "language": "en", "url": "https://arxiv.org/abs/1401.7156", "abstract": "A result of Bennett and Grosse-Erdmann characterizes the weights for which the corresponding weighted Hardy inequality holds on the cone of non-negative, non-increasing sequences and a bound for the best constant is given. In this paper, we improve the bound for $1<p \\leq 2$.", "subjects": "Functional Analysis (math.FA)", "title": "On weighted Hardy inequalities for non-increasing sequences", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9883127406273587, "lm_q2_score": 0.8519528094861981, "lm_q1q2_score": 0.8419958160284825}}
{"url": "https://arxiv.org/abs/1804.02614", "title": "Randomized subspace iteration: Analysis of canonical angles and unitarily invariant norms", "abstract": "This paper is concerned with the analysis of the randomized subspace iteration for the computation of low-rank approximations. We present three different kinds of bounds. First, we derive both bounds for the canonical angles between the exact and the approximate singular subspaces. Second, we derive bounds for the low-rank approximation in any unitarily invariant norm (including the Schatten-p norm). This generalizes the bounds for Spectral and Frobenius norms found in the literature. Third, we present bounds for the accuracy of the singular values. The bounds are structural in that they are applicable to any starting guess, be it random or deterministic, that satisfies some minimal assumptions. Specialized bounds are provided when a Gaussian random matrix is used as the starting guess. Numerical experiments demonstrate the effectiveness of the proposed bounds.", "text": "\\section{Introduction}\nThe computation of low-rank approximations of large-scale matrices is a vital step in many applications in data analysis and scientific computing. These applications include principal component analysis, facial recognition, spectral clustering, model reduction techniques such as proper orthogonal decomposition (POD) and discrete empirical interpolation method (DEIM), approximation algorithms for partial differential  and integral equations. The celebrated Eckart-Young theorem~\\cite{GovL13} says that the optimal low-rank approximation can be obtained by means of the Singular Value Decomposition (SVD); however, computing the full or truncated SVD can be computationally challenging, or even prohibitively expensive for many applications of interest. \n\n\n{\nRandomized algorithms for computing low-rank approximations have become increasingly popular in the last two decades. For example, see the survey papers~\\cite{HMT09,mahoney2011randomized}. Randomized methods have gained in popularity since they are easy to implement, computationally efficient, and numerically robust. Although randomized algorithms tend to have the same asymptotic cost compared to classical methods, they have several advantages that make them suitable for large-scale computing. Specifically,  for  datasets that are too large to fit in memory, randomized algorithms are able to exploit parallel computing efficiently and are efficient in the number of times they access the data.  Randomized algorithms also have excellent numerical robustness and are very reliable in practical applications. \n\n\nWe focus on a specific randomized algorithm known as {\\em randomized subspace iteration}. The main idea of this method is to use random sampling to identify a subspace that approximately captures the range of the matrix. A low-rank approximation to the matrix is then obtained by projecting the matrix onto this subspace. A post-processing step is then performed to compress the low-rank representation to achieve a desired target rank, and a conversion step to obtain an equivalent representation in the desired format (typically, a truncated SVD representation)---both these steps are deterministic. }\n\nMany advances have been made in the analysis of randomized algorithms for low-rank approximations. The analysis typically has two stages: a \\textit{structural, or deterministic stage}, in which minimal assumption about the distribution of the random matrix is made, and a \\textit{probabilistic stage}, in which the distribution of the random matrix is taken into account to derive bounds for expected and tail bounds of the error distribution. As mentioned earlier, existing literature only targets the error in the low-rank representation~\\cite{gu2015subspace,HMT09}. When the low-rank representation is in the SVD format, it is desirable to understand the quality of the approximate subspaces and the individual singular triplets. This paper aims to fill in some of the  missing gaps in the literature by a rigorous analysis of the accuracy of approximate singular values, vectors and subspaces obtained using randomized subspace iteration. This analysis will be beneficial in applications where an analysis beyond the low-rank approximation is desired. Examples include Model Reduction techniques~\\cite{erichson2017randomized,balabanov2018randomized}, Leverage Score computation~\\cite{holodnak2015conditioning}, Spectral Clustering~\\cite{boutsidis2015spectral}, FEAST eigensolvers~\\cite{peter2014feast}, Canonical Correlation Analysis~\\cite{avron2013efficient}.\n\n\n\\subsection{Contributions and overview of paper} \nWe survey the contents and the main contributions of this paper.\n\n\n\\textbf{Canonical angles.} We have developed bounds for \\textit{all} the canonical angles between the spaces spanned by the exact and the approximate singular vectors. Several different flavors of bounds are provided: \n\\begin{enumerate}\n\\item The bounds in \\cref{ssec:angles} relate the canonical angles between the exact and the approximate singular subspaces. Analysis is also provided for unitarily invariant norms of the canonical angles.\n\\item In applications where lower dimensional subspaces are  extracted from the approximate singular subspaces, the bounds in \\cref{ssec:sintheta} quantifies the accuracy in the extraction process. \n\\item \\cref{ssec:sintheta} also presents bounds for the angles between the individual exact and approximate singular vectors, extracted from the appropriate subspaces.\n\\end{enumerate}\nOur bounds suggest that the accuracy of the singular values and vectors, in addition to the low-rank approximations, is high provided (1) singular values decay rapidly beyond the target rank $k$, and (2) the larger the singular value gaps, the higher is the accuracy to be expected. Furthermore, the truncation step to extract the $k$ dimensional subspaces does not significantly lower the accuracy of the subspaces.  \n\n\n\\textbf{Low-rank approximation.} This paper provides the first known analysis of the randomized subspace iteration for an arbitrary unitarily invariant norm, with stronger, specialized results for Schatten-p norms.  Bounds for the special cases of the Schatten-p norm, namely the spectral and Frobenius norms, have already appeared in the literature---our result for the Schatten-p norm recovers these results as special cases.\n \n\n\\textbf{Singular values.} We derive upper and lower bounds on the approximate singular values obtained by the randomized subspace iteration. Similar bounds also appear in~\\cite{gu2015subspace}; however, our proof technique is different. We also present Hoffman-Wielandt type bounds for the accuracy of the singular values. \n\nThe conclusion of the bounds for the low-rank approximations and the singular values are similar to those of the conclusions for the canonical angles. \n\n\\textbf{Generalization of sin theta theorem} The sin theta theorem~\\cite{wedin1983angles} is a well known result in numerical analysis and relates the canonical angles between the true and approximate singular subspaces in the unitarily invariant norms. We derive a generalization of the sin theta theorem that derives bounds for the individual canonical angles between the two subspaces. The sin theta theorem is recovered as a special case. This result {maybe} of independent interest beyond the study of randomized algorithms. \n\n\n\n\\section{Background and preliminaries}\n\\subsection{Notation}\\label{ssec:prelim}\n\n\n{\nDenote the target rank by $k$ and let $1 \\leq k \\leq \\mathsf{rank}\\,(A)$. Let the matrix $A\\in \\mathbb{C}^{m\\times n}$, have the SVD \n\\[A = \\bmat{U_k & U_{\\perp}}\\bmat{\\Sigma_k & \\\\ &\\Sigma_{\\perp} }\\bmat{ V_k^* \\\\ V_\\perp^*}. \\]\nHere, $\\Sigma_k \\in \\mathbb{C}^{k\\times k}$ and $\\Sigma_{\\perp} \\in \\mathbb{C}^{(m-k) \\times (n-k)}$; the columns of $U_k$ and $U_{\\perp}$ are the corresponding left singular vectors, and columns of $V_k$ and $V_{\\perp}$ are the corresponding right singular vectors.  We also denote by $A_k = U_k\\Sigma_kV_k^*$ as the best rank$-k$ approximation to the matrix $A$, in any unitarily invariant norm (for a definition, see below). We also define $A_\\perp = U_\\perp \\Sigma_\\perp V_\\perp^*$ and observe that \n\\[ A = A_k + A_\\perp .   \\]\n  }\n \n \n \\paragraph{Singular values and ratios} Let $\\normtwo{\\cdot}$ denote the spectral norm, so that $\\normtwo{\\Sigma_{\\perp}} = \\sigma_{k+1}$ and $\\normtwo{\\Sigma_k^{-1}} = \\frac{1}{\\sigma_k}$. The singular values of $A$ can be arranged in decreasing order as \n \\[ \\sigma_1 \\geq \\sigma_2 \\geq \\dots \\geq \\sigma_k \\geq \\sigma_{k+1} \\geq \\dots \\geq \\sigma_n.\\]\n For later use, we define the singular value ratios\n \\begin{equation}\n\\gamma_j = \\frac{\\sigma_{k+1}}{\\sigma_j} \\qquad j=1,\\dots,k.\n\\end{equation}\nSince the singular values are monotonically decreasing, the singular value ratios are monotonically increasing, i.e.,  $\\gamma_1 \\leq \\dots \\leq \\gamma_k \\leq 1$.\n\n\n \\paragraph{Norms} We have already defined the spectral norm. The Frobenius norm of a matrix is $\\normf{A} = \\sqrt{\\mathsf{trace}\\,(A^*A)}$. We use the symbol $\\uninorm{\\cdot}$ to denote any unitarily invariant norm, i.e., a norm that satisfies $\\uninorm{QAZ} = \\uninorm{A}$ for unitary matrices $Q,Z$. An example of the unitarily invariant norms is Schatten-p class of norms, defined as the vector $\\ell_p$ norm of the singular values of $A$, i.e.,\n\\[ \\schattenp{A} = \\left( \\sum_{j=1}^{\\min\\{m,n\\}}\\sigma_j^{p}\\right)^{1/p}.\\]\nWith this definition, it can be readily seen that $\\normtwo{A}=  \\uninorm{A}_{\\infty}$ and $\\normf{A} = \\uninorm{A}_{2}$. Another example is the Ky-Fan-$k$ class of norms defined $\\|A \\|_{(k)} = \\sum_{j=1}^k \\sigma_j$ for every $k=1,\\dots,\\min\\{m,n\\}$. Associated with every unitarily invariant norm is a symmetric gauge function acting on the singular values of the matrix that it acts on.\n \n\n\n \\paragraph{Projection matrices} Suppose the matrix $Z$ has full column rank with column space $\\mc{R}\\,(Z)$; $Z^\\dagger$ is a left multiplicative inverse and where ${}^\\dagger$ represents the Moore-Penrose inverse. We define the (orthogonal) projection matrix $\\proj{Z} = ZZ^\\dagger$. {An orthogonal projection matrix is uniquely defined by its range, and $\\mc{R}\\,(\\proj{Z}) = \\mc{R}\\,(Z)$.} For a matrix $Q$ with orthonormal columns, the formula simplifies and $\\proj{Q} = QQ^*$. \n \n\\paragraph{Canonical angles} The separation between subspaces can be measured by the principal or canonical angles. Let $\\mc{M}$ and $\\mc{N}$ be two subspaces of $\\mb{C}^n$, such that $\\dim \\mc{M} = \\ell$, $\\dim\\mc{N} = k$ and $\\ell\\geq k$. Then the principal angles between the subspaces $\\mc{M}$ and $\\mc{N}$ are recursively defined to be the numbers $0 \\leq \\theta_i \\leq \\pi/2$ such that\n\\[ \\cos\\theta_i = \\max_{u \\in \\mc{M}, v \\in \\mc{N} \\\\ \\normtwo{u} = \\normtwo{v} = 1} v^*u = v_i^*u_i, \\qquad i = 1,\\dots,k\\]\nsubject to the constraints $\\normtwo{u_i} = \\normtwo{v_i} = 1$,  and \n\\[u_j^*u, \\quad v_j^*v = 0,  \\qquad j=1,\\dots,i-1.\\] \nThe canonical angles are arranged in increasing order as \n\\[ 0 \\leq \\theta_1 \\leq \\dots \\leq \\theta_k \\leq \\pi/2.\\] \nIt can also be shown that $\\sin\\theta_i$ are also the singular values of $\\proj{\\mc{M}} - \\proj{\\mc{N}}$. \n\nWe denote $\\angle(\\mc{M},\\mc{N})$ to be the canonical angles between subspaces $\\mc{M}$ and $\\mc{N}$.  Let $M$ and $N$ be  matrices with orthonormal columns, which form bases for subspaces $\\mc{M}$ and $\\mc{N}$ respectively. Then, the singular values of  singular values of $(I-MM^*)N$ can be used to compute $\\sin\\angle(\\mc{M},\\mc{N})$ and the singular values of $M^*N$ can be used to compute $\\cos\\angle(\\mc{M},\\mc{N})$~\\cite[Section 3]{bjorck1973numerical}. For ease of notation, in the rest of this paper, we write $\\angle(M,N)$ instead of $\\angle(\\mc{M},\\mc{N})$.\n\n\n\n \n\\subsection{Randomized subspace iteration} The basic version of the randomized subspace iteration is summarized in \\cref{alg:randsvd}. Given a starting guess, denoted by $\\Omega \\in \\mathbb{C}^{n\\times (k+\\rho)}$, the algorithm performs $q$ steps of the randomized subspace iteration to obtain the  matrix $Y$, also known as the ``sketch.'' A thin-QR factorization of $Y$ is performed to obtain $Q$ whose columns form an orthonormal basis for the range of $Y$. The main idea is that, under suitable conditions, the range of $Q$ is a good approximation for the range of $A$.  We obtain a low-rank approximation to $A$ by the projection $\\widehat{A} = QQ^*A$. The rest of the algorithm involves converting this low-rank approximation into the SVD format.\n\\begin{algorithm}\n\\begin{algorithmic}[1]\n\\REQUIRE Matrix $A$, Starting guess $\\Omega \\in \\mathbb{C}^{n\\times (k + \\rho)}$, an integer $q \\geq 0$.\n\\STATE Compute $Y = (AA^*)^qA\\Omega$ \n\\STATE Compute thin QR factorization of $Y$, so that $Y = QR$.\n\\STATE Compute $B = Q^*A$ and its SVD $B = U_{B}\\widehat{\\Sigma} \\widehat{V}^*$. \n\\STATE Compute $\\widehat{U} = QU_{B}$.\n\\RETURN Matrices $\\widehat{U},\\widehat{\\Sigma},\\widehat{V} $ that define $\\widehat{A} \\equiv \\widehat{U}\\widehat{\\Sigma}\\widehat{V}^*$.\n\\end{algorithmic}\n\\caption{Idealized version of Subspace iteration for Singular Value Decomposition}\n\\label{alg:randsvd}\n\\end{algorithm}\n The algorithm to compute an approximate singular value decomposition, given starting guess $\\Omega \\in \\mb{C}^{n\\times (k+\\rho)}$ is summarized in~\\cref{alg:randsvd}. We say that this is  an idealized version, since the algorithm can behave poorly in the presence of round-off errors. A practical implementation of this algorithm alternates the QR factorization with matrix-vector products (matvecs) involving $A$; for more details regarding the implementation, the reader is referred to~\\cite{saad2011numerical,HMT09}. In~\\cref{alg:randsvd}, the output $$\\widehat{A} \\equiv QQ^*A =  \\widehat{U}\\widehat{\\Sigma}\\widehat{V}^*$$\nmay have a larger rank than (or equal to) $k$.  If a rank-$k$ approximation to $A$ is desired, then it can be obtained by discarding the $\\rho$ smallest singular values of $\\widehat{A}$. We denote this low-rank representation by \n\\[ \\widehat{A}_k = \\widehat{U}_k \\widehat{\\Sigma}_k \\widehat{V}_k^*. \\]\nThis is summarized in~\\cref{alg:truncation}. \n \n \\begin{algorithm}[!ht]\n \\begin{algorithmic}[1]\n \\REQUIRE Matrix $A\\in \\mathbb{C}^{m\\times n}$ and $Q \\in \\mathbb{C}^{m\\times (k+\\rho)}$. Target rank $1 \\leq k \\mathsf{rank}\\,(A)$. \n \\STATE Form matrix $B = Q^*A$. \n \\STATE Compute the truncated SVD representation $B_k = \\widehat{U}_{B,k} \\widehat{\\Sigma}_k \\widehat{V}_k^*$.\n \\STATE Form $\\widehat{U}_k = Q\\widehat{U}_{B,k}$\n \\RETURN Matrices $\\widehat{U}_k,\\widehat{\\Sigma}_k,\\widehat{V}_k$  such that $\\widehat{A}_k = \\widehat{U}_k\\widehat{\\Sigma}_k\\widehat{V}_k^*$.\n \\end{algorithmic}\n \\caption{Truncated SVD of $\\widehat{A} = QQ^*A$}\n \\label{alg:truncation}\n \\end{algorithm}\n\n\\noindent  {Before we state the assumptions needed for our analysis, we introduce the following notation. The matrix $V^*\\Omega$ captures the influence of the starting guess on the right singular matrix $V$. Partition this matrix as\n\\begin{equation}\\label{eqn:omegadef} V^*\\Omega = \\bmat{V_{k}^*\\Omega \\\\ V_{\\perp}^*\\Omega} = \\bmat{\\Omega_1 \\\\ \\Omega_2},\\end{equation}\nwhere $\\Omega_1 = V_k^*\\Omega \\in \\mathbb{C}^{k\\times (k+\\rho)}$ and $\\Omega_2 = V_\\perp^*\\Omega \\in \\mathbb{C}^{(n-k)\\times (k+\\rho)}$. \nAs was mentioned earlier, we assume that the  target rank $k$ satisfies $1 \\leq k \\leq \\mathsf{rank}\\,(A)$. Additionally, the following assumptions will be required for our analysis.\n\\begin{assumption}\\label{ass:main} \nLet $\\Omega_1 \\in \\mathbb{C}^{k\\times (k+\\rho)}$ be defined as above. We assume that \n\\begin{equation}\\label{eqn:omega1} \\mathsf{rank}\\,(\\Omega_1) = k.  \\end{equation}  \n\\item The singular value gap at index $k$ is inversely proportional to the singular value ratio\n\\begin{equation}\n\\gamma_k = \\normtwo{\\Sigma_\\perp}\\normtwo{\\Sigma_k^{-1}}= \\frac{\\sigma_{k+1}}{\\sigma_k} < 1.\n\\end{equation}\n\\end{assumption}\nThe first assumption guarantees that the starting guess $\\Omega$ has a significant influence over the right singular vectors, whereas the second assumption ensures that the $k$ dimensional subspace $\\mc{R}\\,(U_k)$ is well defined. In practice, it is highly desirable that $\\gamma_k \\ll 1$, which ensures that there is a large singular value gap.}\n\n\\section{Accuracy of singular vectors}\\label{sec:angles}\n\nWe want to understand how well $\\mc{R}\\, (\\widehat{U})$ approximates $\\mc{R}\\,(U_k)$, measured in terms of the canonical angles between the subspaces. To this end, abbreviate the subspace angles between $\\widehat{U} \\in \\mathbb{C}^{m\\times \\ell}$ and $U_k \\in \\mathbb{C}^{m \\times k}$ as $\\theta_1,\\dots,\\theta_k$. Similarly, denote the angles between $\\widehat{V} \\in \\mathbb{C}^{n\\times \\ell}$ and $V_k \\in \\mathbb{C}^{n \\times k}$ by $\\nu_1,\\dots,\\nu_k$. We are also interested in obtaining bounds for the canonical angles $\\angle(U_k,\\widehat{U}_k)$ and $\\angle(V_k,\\widehat{V}_k)$. To distinguish these angles from $\\angle(U_k,\\widehat{U})$ and $\\angle(V_k,\\widehat{V})$, we call them $\\theta_j'$ and $\\nu_j'$ for $j=1,\\dots,k$. \n\\subsection{Bounds for canonical angles}\\label{ssec:angles}\n\nOur first result derives bounds for the canonical angles $\\angle(U_k,\\widehat{U})$. The analysis is based on the perturbation of projectors and the tools used here are similar to~\\cite{HMT09}. \n\n\\begin{theorem}\\label{thm:u}\n Let $\\widehat{U}$ and $\\widehat{V}$ be obtained from~\\cref{alg:randsvd}. With~\\cref{ass:main}, the canonical angles $\\theta_j$ and $\\nu_j$ satisfy\n\\[ \\sin\\theta_j\\leq \\frac{\\gamma_{j}^{2q+1} \\normtwo{\\Omega_2\\Omega_1^\\dagger}}{\\sqrt{1 + \\gamma_{j}^{4q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2}}  \\qquad \\sin\\nu_j \\leq \\frac{\\gamma_{j}^{2q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}}{\\sqrt{1 + \\gamma_{j}^{4q+4} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2}}   \\]\n for $j=1,\\dots,k$. \n\\end{theorem}•  \n{\nThis theorem has several interesting features worth pointing out. First, if the matrix has exact rank $k$, then all of the canonical angles are uniformly equal to zero; that is, the randomized subspace iteration identifies the subspace exactly.  On the other hand, when $\\gamma_k$ is very close to $1$, the subspaces may not be well-defined and {may be} difficult to identify. In practice, it is highly desirable that $\\gamma_k \\ll 1$, so that the angles are captured accurately. \n\n\n\nSecond, the bounds for the canonical angles show explicit dependence on the singular value ratios $\\gamma_j$. In particular, the canonical angles $\\theta_j$ and $\\nu_j$ converge to zero quadratically but at different rates depending on the singular value ratios.  Specifically, the smaller the singular value ratio, smaller the canonical angles. \n\n\n\nThird, the term $\\normtwo{\\Omega_2\\Omega_1^\\dagger}$ can be written in terms of the right singular vector matrix $V$ and the starting guess $\\Omega$ as \n\\[ \\normtwo{\\Omega_2\\Omega_1^\\dagger} = \\normtwo{ (V_{\\perp}^*\\Omega)(V_k^*\\Omega)^\\dagger }.\\]\nWhen the columns of $\\Omega$ is linearly independent, this quantity is nothing but the tangent of the largest canonical angle between $\\mc{R}\\,(V_k)$ and $\\mc{R}\\,(\\Omega)$. This term appears frequently in randomized linear algebra and can be interpreted as  a measure of the subspace overlap between the starting guess and the right singular vectors. In the ideal case, $\\Omega$ contains the singular vectors in $V_k$. \nA discussion of the meaning and interpretation of this term, is provided in~\\cite[Section 2.5]{drineas2017structural}. In particular, when $\\Omega$ is a Gaussian random matrix, $\\normtwo{\\Omega_2\\Omega_1^\\dagger}$ is roughly on the order of $\\sqrt{(n-k)k}$. \n\nFourth, the influence of $\\normtwo{\\Omega_2\\Omega_1^\\dagger}$ is subdued by the singular value ratios $\\gamma_j^{2q+1}$.  With sufficiently large number of iterations $q$, the canonical angles are smaller than a user-defined tolerance. Rigorous bounds for the requisite number of iterations are provided in~\\cref{ssec:prob}.\n\n\nLastly, the bounds for the canonical angles  $\\theta_j$ are smaller than $\\nu_j$ because the latter contains an additional power of $\\gamma_j$. The reason for this higher accuracy is as follows: the columns of $\\widehat{V}$ are the right  singular vectors of $Q^*A$. Therefore, the multiplication step with $Q$ amounts to an additional step of subspace iteration and gives the extra factor. \n}\n\n\n\n\\begin{remark}\n\\cref{thm:u} gives the sine of the canonical angles; these bounds can also be used to obtain upper bounds for the tangents and lower bounds for the cosines. \n With the same assumptions and notation as in~\\cref{thm:u}, the relationship between the tangent and sine implies \n \\[\\tan\\theta_j \\leq {\\gamma_{j}^{2q+1} \\normtwo{\\Omega_2\\Omega_1^\\dagger}} \\qquad \\tan\\nu_j \\leq {\\gamma_{j}^{2q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}} \\]\n for $j=1,\\dots,k$. Lower bounds for cosine of the canonical angles follow similarly.\n\\end{remark}\n\n\n\\paragraph{Unitarily invariant norms}  The following result derives bounds for the canonical angles in any unitarily invariant norm, in contrast to~\\cref{thm:u} which bounds the individual canonical angles. \n\n\\begin{theorem}\\label{thm:theta2F} Let the approximate singular vectors $\\widehat{U}$ and $\\widehat{V}$ for a matrix $A$ be computed according to~\\cref{alg:randsvd}. Under~\\cref{ass:main}, for every unitarily invariant norm, \n\\begin{equation}\n\\begin{aligned}\n \\uninorm{\\sin\\angle(U_k,\\widehat{U}) } \\leq & \\>  \\gamma_k^{2q} \\frac{ \\uninorm{ \\Sigma_\\perp}}{\\sigma_k} \\normtwo{\\Omega_2\\Omega_1^\\dagger}, \\\\ \n \\uninorm{\\sin\\angle(V_k,\\widehat{V}) } \\leq & \\> \\gamma_k^{2q+1} \\frac{\\uninorm{ \\Sigma_\\perp}}{\\sigma_k} \\normtwo{\\Omega_2\\Omega_1^\\dagger}. \n\\end{aligned}\n\\end{equation}\n\\end{theorem}\n\n\nThe interpretation of this theorem is similar to that of~\\cref{thm:u}. The connection between the two theorems follows from the identity $\\sin\\theta_k = \\|\\sin\\angle(U_k,\\widehat{U}) \\|_{2}$. If we specialize the result in~\\cref{thm:theta2F} to the spectral norm, then it is clear that this result weaker than the bound in~\\cref{thm:u}. \n\n\\subsection{Extraction of $k$-dimensional subspaces}\\label{ssec:sintheta}\nIn the previous subsection, the columns of $\\widehat{U}$ and $\\widehat{V}$ spanned $\\ell = k +\\rho$ dimensional subspaces. Many applications, however, require the extraction of $k$ dimensional singular subspaces from the low-rank approximation $\\widehat{A} \\equiv QQ^*A$. One way to extract the appropriate subspaces is to first compute the optimal rank-$k$ truncation of $\\widehat{A}$, denoted by $\\widehat{A}_k$. The singular vectors of $\\widehat{A}_k$, denoted by $\\widehat{U}_k$ and $\\widehat{V}_k$, are then used instead of $\\widehat{U}$ and $\\widehat{V}$. See~\\cref{alg:truncation}, for details regarding implementation. The bounds derived in the previous subsection are not directly applicable since~\\cite[Corollary 10]{ye2016schubert} says \n\\[ \\theta_j \\leq \\theta_j'  \\qquad \\nu_j \\leq \\nu_j' \\qquad j=1,\\dots,k.\\] \nTo understand how much additional error is incurred during this extraction process, we present several results. The important conclusion of all these results is that the accuracy of the extracted subspaces of dimension $k$ is comparable to the accuracy of the $k+\\rho$ dimensional subspace provided the singular values are sufficiently well separated.\n\n\n\nThe approach we take is different from that of the previous section. The starting point of our analysis is the well-known sin theta theorem for singular subspaces~\\cite{wedin1983angles}. Let $A,\\widehat{A}$ be two matrices of conformal dimensions. Assuming that  \n\\begin{equation}\\label{eqn:gap}\n\\zeta \\equiv \\sigma_k(A) - \\sigma_{k+1}(\\widehat{A}) > 0,\n\\end{equation}\nwe have\n\\begin{equation}\\label{eqn:sintheta} \\max\\left\\{ \\uninorm{\\sin \\angle(U_k,\\widehat{U}_k)} ,\\uninorm{\\sin\\angle(V_k,\\widehat{V}_k)} \\right\\} \\leq \\frac{\\max\\{\\uninorm{E_{12}},\\uninorm{E_{21}}\\}}{\\zeta},\n\\end{equation}\n where  the two matrices $E_{12}$ and $E_{21}$ are \n\\begin{equation} \\label{eqn:e1}\n\\begin{aligned}\nE_{12}  = & \\>(I-\\proj{\\widehat{U}_k}) (A-\\widehat{A}) \\proj{V_k}\\\\\nE_{21}  = &  \\> \\proj{{U}_k} (A-\\widehat{A}) (I-\\proj{\\widehat{V}_k}).\n\\end{aligned} \n\\end{equation}\n\nHowever, this version of the sin theta theorem does not provide us with a way to obtain bounds for the individual canonical angles. To this end, we first present a new generalization of the sin theta theorem. \n\\begin{theorem}\\label{thm:gensintheta} Let $A \\in \\mathbb{C}^{m\\times n}$ with $\\mathsf{rank}\\,(A) \\geq k$ and let $\\widehat{A}$ be the perturbed matrix with same dimensions. Suppose the singular value gap satisfies~\\cref{eqn:gap}. \nLet $\\widehat{A}_k = \\widehat{U}_k\\widehat{\\Sigma}_k\\widehat{V}_k^*$ be the truncated SVD of $\\widehat{A}$. Then \n\\[ \\max \\{ \\sin \\theta_j' , \\sin \\nu_j' \\} \\leq \\frac{\\sigma_k(A)}{\\sigma_{j}(A)} \\max\\{\\sin\\theta_k',\\sin\\nu_k'\\} \\qquad j=1,\\dots,k.\\]\n\\end{theorem}\nThis theorem states that the sine of the canonical angles $\\sin \\theta_j'$ are bounded by  $\\sin \\theta_k' $ up to a multiplicative factor, which is at most $1$.\n\n\nOur main result provides the following bounds for canonical angles between the exact and the approximate singular subspaces, when both the subspaces have the same dimension. The proof involves simplifying every term in~\\cref{eqn:e1}.\n\\begin{theorem}\\label{thm:sintheta} \nLet $\\widehat{U}$ and $\\widehat{V}$ be obtained from~\\cref{alg:randsvd}, and matrices $\\widehat{U}_k$ and $\\widehat{V}_k$ from~\\cref{alg:truncation}. Under~\\cref{ass:main}, \n\\begin{itemize}\n\\item for every unitarily invariant norm \n\\[  \\max\\left\\{ \\uninorm{ \\sin \\angle (U_k,\\widehat{U}_k) } ,\\uninorm{\\sin\\angle(V_k,\\widehat{V}_k)} \\right\\} \\leq \\phi\\frac{\\gamma_k^{2q}}{1-\\gamma_k} \\frac{\\uninorm{\\Sigma_\\perp}}{\\sigma_k} \\normtwo{\\Omega_2\\Omega_1^\\dagger}.\\]\nThe factor $\\phi$ takes different values depending on the specific norm used. For an arbitrary unitarily invariant norm, we have  $\\phi= \\sqrt{2}$, whereas for the spectral and Frobenius norms, we have $\\phi = 1$.\n\\item  canonical angles $\\theta_j'$ and $\\nu_j'$ satisfy\n\\[\\max \\{ \\sin \\theta_j' , \\sin \\nu_j' \\} \\leq \\gamma_{j} \\frac{\\gamma_k^{2q}}{1-\\gamma_k}\\normtwo{\\Omega_2\\Omega_1^\\dagger} \\qquad j=1,\\dots,k.\\]\n\\end{itemize}• \n\\end{theorem}\nThe interpretation of this theorem is: (1) as the number of iterations $q$ increase,  the largest canonical angle converges to $0$ quadratically, and (2) a larger singular value gap means that the subspace is computed more accurately.  Comparing this result with~\\cref{thm:theta2F}, we see that the upper bound in~\\cref{thm:sintheta} has additional factors which depend on the specific norm used. For an arbitrary unitarily invariant norm, there is an additional factor $\\max\\{1,\\sqrt{2}\\gamma_k\\}/(1-\\gamma_k)$.  For the spectral and Frobenius norms, the additional factor is $1/(1-\\gamma_k)$.  Both factors are greater than $1$, suggesting that the truncation process can introduce additional error. The additional factor is also independent of the number of iterations $q$, suggesting that it is a one-time price to be paid for the extraction process. The  bound is devastating when $\\gamma_k \\approx 1$, but this also means that the subspaces may not be well-defined.\n\n\\paragraph{Individual singular vectors}\nThe previous results give insight into the accuracy measured using the canonical angles between the exact and approximate singular subspaces. When individual singular vectors need to be extracted, does the extraction process introduce additional error? The following result quantifies the accuracy of the extraction process.\n\\begin{theorem}\\label{t_singleuv}\nLet the approximate singular vectors $\\widehat{U}$ and $\\widehat{V}$ be computed according to~\\cref{alg:randsvd}. With~\\cref{ass:main}, we have the following inequalities  \n\\begin{equation}\\label{eqn:singleuv}\n \\sin \\angle(u_j,\\widehat{U})\\leq   {\\gamma_{j}^{2q+1} \\normtwo{\\Omega_2\\Omega_1^\\dagger}} \\qquad \n\\sin\\angle(v_j,\\widehat{V}) \\leq  {\\gamma_{j}^{2q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}}\n\\end{equation}\nfor $j=1,\\dots k$. Denote the approximate singular triplets $(\\hat\\sigma_j,\\hat{u}_j,\\hat{v}_j)$ for $j=1,\\dots,k$. Under~\\cref{ass:main}\n\\begin{equation}\\label{eqn:svs}\n \\max\\left\\{\\sin\\angle(u_j,\\hat{u}_j),\\sin\\angle(v_j,\\hat{v}_j) \\right\\} \\leq \\sqrt{1 + 2\\frac{\\tilde{\\gamma}^2}{\\tilde\\delta^2}} \\, \\gamma_j^{2q+1}\\normtwo{\\Omega_2\\Omega_1^\\dagger}.\n\\end{equation}\nHere, $\\tilde\\gamma^2 \\equiv \\normtwo{\\Sigma_\\perp}^2 + \\normtwo{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger}^2 $ and $\\tilde\\delta \\equiv \\min\\{\\min_{\\tilde{\\sigma}_i\\neq \\tilde\\sigma_j}\\{ |\\sigma_j - \\tilde\\sigma_i| ,\\sigma_j\\} \\} $.\n\\end{theorem}• \nThe first result bounds the angles between the exact singular vector and the corresponding approximate singular subspaces. The second result compares the angles of the exact and the approximate singular vectors. This result also says that the extraction process does not adversely increase the error in the singular subspaces, provided the singular values are well-separated. \n \nThe convergence of the individual singular vectors tell a similar story to that of~\\cref{thm:u}. The singular vectors corresponding to the largest singular values converge earlier than the singular vectors corresponding to the smaller singular vectors. This is a consequence of the fact that the singular value ratios are non-decreasing. \n\n\\subsection{Comparison with other bounds}\n\nThe subspace iteration dates to a 1957 paper by Bauer~\\cite{Bauer1957} for eigenvalue problems. The analysis of the subspace iteration has also been well-established, for example, we refer to~\\cite[Chapter 14]{Par80}. {Randomized subspace iteration has attracted a lot of attention in the last two decades, with a special emphasis on quantifying the influence of the starting guess $\\Omega$. In particular, recent research has focused on the choice of the distribution and the effect of the oversampling parameter $\\rho$.  The effect of randomized subspace iteration on the accuracy of singular vectors was studied in the context of spectral clustering in~\\cite{boutsidis2015spectral}. However, the authors made the rather strong assumption that $\\Omega \\in \\mathbb{R}^{n\\times k}$, which amounts to setting the oversampling parameter $\\rho = 0$. This is a strong requirement since~\\cref{ass:main} now requires $\\Omega_1$ to be invertible.  The authors were able to show (in our notation)\n\\[ \\normtwo{ \\sin \\angle (U_k,\\widehat{U}_k)} \\leq \\frac{\\gamma_k^{2q+1}\\normtwo{\\Omega_2\\Omega_1^{-1}}}{\\sqrt{1 + \\gamma_k^{4q+2}\\normtwo{\\Omega_2\\Omega_1^{-1}}^2}}.\\]\nNotice that this bound coincides with~\\cref{thm:u} (for $\\sin\\theta_k$) when $\\rho = 0$. Our results provide bounds for the right singular vectors as well as all the canonical angles.\n\n\nLet us return to this assumption that $\\mathsf{rank}\\,(\\Omega_1)$. When $\\Omega$ is standard Gaussian matrix,~\\cite[Theorem 3.3]{sankar2006smoothed} says\n $$\\|\\Omega_1^{-1}\\|_2 \\leq \\frac{2.35 \\sqrt{k}}{\\delta}$$\n with probability at least $1-\\delta$. For a small probability of failure $0 <\\delta < 1$, this bound can be devastating. By contrast, if we let $\\Omega_1 \\in \\mathbb{C}^{k\\times (k+\\rho)}$ with $\\rho \\geq 2$, and still suppose that $\\Omega$  is a Gaussian random matrix. Then, with probability at least $1-\\delta$~\\cite[Proposition 10.4]{HMT09} says\n$$\\normtwo{\\Omega_1^\\dagger} \\leq e\\frac{\\sqrt{k+\\rho}}{\\rho} \\left(\\frac{1}{\\delta}\\right)^{1/(\\rho+1)}. $$ \nIt is clear that when the random matrix is Gaussian, oversampling  has an impact on the accuracy of the randomized subspace iteration. Specifically, larger the oversampling, the more accurate is the subspace.  }\n\n\nOversampling plays a bigger role for random matrices that have different distributions than Gaussian. When $\\Omega$ is generated from the subsampled randomized Hadamard transform (SRHT), or Rademacher distributions, a more aggressive form of oversampling $\\ell \\sim k\\log k$ is necessary to ensure that $\\mathsf{rank}\\,(\\Omega_1) = k$. Therefore, by allowing for oversampling, our bounds are applicable to starting guesses that are not restricted to Gaussian random matrices. Not only that, our bounds are also informative for matrices with decaying singular values and significant singular value gap. \n\n\nA recent paper by Nakatsukasa~\\cite{nakatsukasa2017accuracy} considered the issue of accuracy of extracting singular subspaces for general projection-based approximation methods. In our notation, these refer to relating bounds for $\\angle(U_k,\\widehat{U})$ to $\\angle(U_k,\\widehat{U}_k)$.  Our bounds for the canonical angles appear to be tighter than the result implied by~\\cite[Corollary 1]{nakatsukasa2017accuracy}. This may be because the analysis was applicable to arbitrary subspace projections, whereas ours is specialized to randomized subspace iteration; we do not go into a detailed comparison here. Furthermore, our analysis is able to bound the individual canonical angles which is missing in~\\cite{nakatsukasa2017accuracy}. \n\n\n\n\n\\subsection{Probabilistic bounds}\\label{ssec:prob} Thus far, we\nhave not made specific assumptions on the matrix $\\Omega$, as long as it\nsatisfies $\\mathsf{rank}\\,(\\Omega_1) = k$. In particular, $\\Omega$ need not be even be random, and\nmay be deterministic. However, more can be said about the bounds when $\\Omega$ is random\nis drawn from a specific distribution.\n\nIn many applications, the matrix $\\Omega \\in \\mathbb{R}^{n\\times (k+\\rho)}$ is taken to be the standard Gaussian random matrix. That is, the entries of $\\Omega$ are i.i.d.\\ $\\mc{N}(0,1)$ random variables. Here we derive a few probabilistic results that provide insight into the accuracy of the subspaces. Let $\\rho \\geq 2$ and define the constant\n\\begin{equation}\n\\label{eqn:ce}\n C_e = \\sqrt{\\frac{k}{\\rho-1}} + \\frac{e\\sqrt{(k+\\rho)(n-k)}}{\\rho}\n \\end{equation}\nand for $0 < \\delta < 1$ define the constant\n\\begin{equation}\n\\label{eqn:cd}\nC_d = \\frac{e\\sqrt{k+\\rho}}{\\rho+1}\\left(\\frac{2}{\\delta}\\right)^{1/(\\rho+1)} \\left( \\sqrt{n-k} + \\sqrt{k+\\rho} + \\sqrt{2\\log\\frac{2}{\\delta}}\\right).\n\\end{equation}\n\\begin{theorem}[Probabilistic bounds]\\label{thm:can_gauss} Let $\\Omega \\in \\mb{R}^{n\\times (k+\\rho)}$ be a standard  Gaussian random matrix with $\\rho \\geq 2$. Assume that the singular value ratio $\\gamma_k < 1$. Let $\\widehat{U}$ and $\\widehat{V}$ be obtained from~\\cref{alg:randsvd}. For $j=1,\\dots,k$, the expected value of the canonical angles satisfy\n\\[ \\mb{E}\\, \\left[\\sin\\theta_j \\right] \\leq \\frac{\\gamma_{j}^{2q+1}C_e}{\\sqrt{1 + \\gamma_{j}^{4q+2}C_e^2} } \\qquad \\mb{E} \\, \\left[\\sin\\nu_j \\right] \\leq \\frac{\\gamma_{j}^{2q+2}C_e}{\\sqrt{1 + \\gamma_{j}^{4q+4}C_e^2} }.\\]\nLet $0 < \\delta < 1$ be a user defined failure tolerance. With probability, at least $1-\\delta$, the following inequalities hold independently for $j=1,\\dots,k$\n\\[  \\sin\\theta_{j} \\leq \\frac{\\gamma_{j+1}^{2q+1}C_d}{\\sqrt{1 + \\gamma_{j}^{4q+2}C_d^2} } \\qquad \\sin\\nu_j \\leq \\frac{\\gamma_{j}^{2q+2}C_d}{\\sqrt{1 + \\gamma_{j}^{4q+4}C_d^2} }.\\]\n \\end{theorem}\n\nThe main message of theorem can be seen from the following bound on the number of subspace iterations $q$.  Specifically, suppose $0 < \\epsilon < 1$, and the number of subspace iterations $q$ we take satisfies \n\\[ q \\geq \\frac12 \\left( \\frac{\\log \\epsilon/C_e}{\\log \\gamma_k} -1\\right),\\]\nthen $  \\mb{E}\\, \\sin \\theta_j \\leq \\mc{O}(\\epsilon^2)$ for $j=1,\\dots,k$.\n\n\n Several extensions of these results are possible. First, following the proof technique of~\\cref{thm:can_gauss}, we can extend the probabilistic analysis to~\\cref{thm:theta2F,t_singleuv} as well. Second, following the strategy in~\\cite{HMT09}, the probabilistic results can be extended to other distributions. However, we will not pursue these extensions here.\n\n\n\n\\section{Low-rank approximation and Singular values}\\label{sec:aux}\n\nIn this section, we provide several structural bounds for the accuracy of the low-rank approximation and the accuracy of the singular values.\n\n\\subsection{Low-rank approximation} Several results are available for\nestimating the error in the low-rank approximation $A \\approx QQ^*A$ in the\nspectral and Frobenius norms, when the matrix $Q$ is obtained from the\nrandomized subspace\niteration~\\cite{HMT09,gu2015subspace,zhang2016randomized}. As was mentioned\nearlier, the spectral and Frobenius norms are special cases of the Schatten-p\nnorm, which are examples of unitarily invariant norms. \n\n\nHere we present the first known analysis of randomized subspace iteration in a unitarily invariant norm. \n\\begin{theorem}\\label{thm:lowrank_schatten}\nLet $\\widehat{A} \\in \\mathbb{C}^{m\\times n}$ be computed using~\\cref{alg:randsvd}. Under~\\cref{ass:main}, the following inequalities hold in every unitarily invariant norm \n\\begin{align}\\label{eqn:nnorm}\n \\uninorm{(I-QQ^*)A} \\leq & \\> \\uninorm{\\Sigma_\\perp}+ \\gamma_k^{2q}\\uninorm{\\Sigma_\\perp\\Omega_2 \\Omega_1^\\dagger } \\\\ \\label{eqn:nnorm_rankr}\n\\uninorm{(I-QQ^*)A_k } \\leq & \\>  \\gamma_k^{2q}\\uninorm{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger}.\n\\end{align}•\nLet $B =Q^*A$, and let $B_k$ be its best rank$-k$ approximation. If $A$ is approximated using $QB_k$, then the error in the low-rank approximation is\n\\begin{equation}\\label{eqn:nnorm_rankr2}\n\\uninorm{A-QB_k} \\leq  \\> \\left( 1+ \\frac{\\sigma_1}{\\sigma_k}\\frac{\\phi\\gamma_k^{2q}}{1-\\gamma_k} \\normtwo{\\Omega_2 \\Omega_1^\\dagger} \\right)\\uninorm{\\Sigma_\\perp}.\\end{equation}\n As in~\\cref{thm:sintheta}, $\\phi = 1$ for spectral and Frobenius norms, and $\\sqrt{2}$ for an arbitrary unitarily invariant norm. \n\\end{theorem}\nIn this theorem, as the number of iterations $q\\rightarrow \\infty$, the error in the low-rank approximation goes to zero.\n\n\nWe present a variant of the error in the low-rank approximation for the special case that a Schatten-p norm is used.  The proof for the special case of the Frobenius norm was provided in~\\cite{zhang2016randomized}. \n\\begin{theorem}\\label{thm:lowrank}\nLet $\\widehat{A}$ be computed using~\\cref{alg:randsvd}. Under~\\cref{ass:main}, we have\n\\begin{equation}\\label{eqn:2norm}\n\\schattenp{ (I-QQ^*)A }^2 \\leq \\schattenp{\\Sigma_\\perp}^2 + \\gamma_k^{4q}\\schattenp{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger}^2.\n\\end{equation}•\n\\end{theorem}\n\nThe error bound in~\\cref{thm:lowrank_schatten} is weaker than~\\cref{thm:lowrank} for the Schatten-p norm  since for $\\alpha,\\beta \\geq\n0$, we have $\\sqrt{\\alpha^2 + \\beta^2} \\leq \\alpha + \\beta$. More generally,~\\cref{thm:lowrank} is applicable to any unitarily invariant norm that is\nalso a Q-norm~\\cite[Definition IV.2.9]{bhatia1997matrix}. A unitarily invariant norm $\\uninorm{\\cdot}_Q$ is a $Q$-norm, if there exists\nanother unitarily invariant norm $\\uninorm{\\cdot}_a$ such that $\\uninorm{A}_Q^2 =\n\\uninorm{A^*A}_a$. Note that the Schatten-p norms satisfy this property for $p\n\\geq 2$, since $\\schattenp{A}^2 = \\uninorm{A^*A}_{p/2}$.\n\n\n\\subsection{Accuracy of singular values} How are the singular values of $A$\nrelated to the singular values of $\\widehat{A}$? We now present a result that\nquantifies the accuracy of the individual singular values. This result is\nsimilar to~\\cite[Theorem 4.3]{gu2015subspace}. Our proof techniques are\nsubstantially different. We make extensive use of the Cauchy interlacing\ntheorem and the multiplicative singular value inequalities~\\cref{eqn:singprod}.\n \n\\begin{theorem}\\label{thm:sigma}\nLet $\\widehat{A} = \\widehat{U}\\widehat{\\Sigma}\\widehat{V}^*$ be computed using~\\cref{alg:randsvd}. Under~\\cref{ass:main}, the\napproximate singular values $\\sigma_j(\\widehat{A})$ satisfy for $j=1,\\dots,k$ \n\\[ \\sigma_j(A) \\geq \\sigma_j(\\widehat{A}) \\geq \\frac{\\sigma_j(A)}{\\sqrt{1 + \\gamma^{4q+2}_{j}\\normtwo{\\Omega_2\\Omega_1^\\dagger}^2 }}.\\]\n\\end{theorem}•\nIt can be readily seen that the large singular values are computed more accurately since the singular value ratio corresponding to larger singular values is smaller. \n\nRather than quantify the accuracy of the individual singular values, the next results are of the Hoffman-Wielandt type and account for all the singular values together. Define the two matrices of conformal sizes\n\\[ \\Sigma = \\bmat{\\Sigma_k \\\\ & \\Sigma_\\perp} \\qquad \\Sigma' = \\bmat{\\widehat{\\Sigma} \\\\ & 0}.\\]\nUnder~\\cref{ass:main}, the error in the singular values satisfies\n\\begin{equation}\n \\uninorm{\\Sigma-\\Sigma'} \\leq \\uninorm{\\Sigma_\\perp}  + \\gamma_k^{2q} \\uninorm{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger} .\n\\end{equation}\nThe proof combines~\\cite[III.6.13]{bhatia1997matrix} with~\\cref{thm:lowrank_schatten}. For the Schatten-p norm, with $p\\geq 2$, we can derive the bound\n\\begin{equation}\\schattenp{\\Sigma-\\Sigma'} \\leq \\sqrt{\\schattenp{\\Sigma_\\perp}^2  + \\gamma_k^{4q} \\schattenp{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger}^2} .\n\\end{equation}\nThe proof is similar, and is therefore omitted.\n\n\\section{Proofs}\\label{sec:proofs}\nWe recall some results here that will be useful in our analysis, see~\\cite[Section 7.7]{HoJ13} for proofs.  Let $M,N$ be Hermitian positive definite. The notation $M \\preceq N$ means $N-M $ is positive semi-definite and it defines a partial ordering on the set of Hermitian matrices. Clearly, this also implies $I -N \\preceq I -M$. The partial order is preserved under the conjugation rule. That is\n\\[ SMS^* \\preceq SNS^* \\qquad \\forall \\> S \\in \\mb{C}^{m\\times n}. \\]\nWeyl's theorem implies that the eigenvalues satisfy  $\\lambda_j(M) \\leq \\lambda_j(N)$ for all $j=1,\\dots,n$. If additionally, $M,N$ are both positive semidefinite then $M^{1/2} \\preceq N^{1/2}$~\\cite[Proposition V.1.8]{bhatia1997matrix} and $(I+N)^{-1} \\preceq (I+M)^{-1}$. \n \n \n  \\paragraph{Singular value inequalities} Let $A,B \\in \\mathbb{C}^{m\\times n}$. For all $i,j$ such that $1 \\leq i,j \\leq \\min\\{m,n\\}$ and $i+j-1\\leq \\min\\{m,n\\}$, the following singular value inequalities hold for the sum $A+B$~\\cite[Equation 7.3.13]{HoJ13}\n \\begin{equation}\n\\sigma_{i+j-1}(A+B) \\leq \\sigma_i(A) + \\sigma_j(B),\n\\end{equation}\nand product $AB^*$~\\cite[Equation (7.3.14)]{HoJ13}\n\\begin{equation}\\label{eqn:singprod}\n\\sigma_{i+j-1}(AB^*)\\leq \\sigma_i(A)\\sigma_j(B).\n\\end{equation}•\n A useful corollary of these results is that $\\sigma_i(A+B) \\leq \\sigma_i (A) + \\sigma_1(B)$ and $\\sigma_i(AB^*) \\leq \\sigma_i(A) \\sigma_1(B)$ for $i=1,\\dots,\\min\\{m,n\\}$.\n\n\n \n\\paragraph{Unitarily invariant norms} {It is useful to recall some properties of the unitarily invariant norms. Every unitarily invariant norm $\\uninorm{\\cdot}$ on $\\mathbb{C}^n$ is associated with a symmetric gauge function on $\\mathbb{R}^n$. The  $\\uninorm{\\cdot}$  satisfies $\\uninorm{M} = \\uninorm{(M^*M)^{1/2}}$, since both matrices have the same nonzero singular values. The following inequality for unitarily invariant norms, also known as strong sub-multiplicativity, will be useful~\\cite[(IV.40)]{bhatia1997matrix}  \n$$\\uninorm{ABC} \\leq \\normtwo{A}\\normtwo{C}\\uninorm{B}.$$\n\n\nWe will need the following lemma\n \\begin{lemma}\\label{lemma:schatten}\n Let $A, B, D \\in \\mathbb{C}^{n\\times n} $ such that  $A,B$ Hermitian and  $0 \\preceq A \\preceq B$, then \n \\[ \\uninorm{(D^*AD)^{1/2}} \\leq \\uninorm{(D^*BD)^{1/2}}.\\]\n \\end{lemma}\n \n \\begin{proof}\n Combining the properties of the partial ordering, the eigenvalues of the scaled matrices satisfy $\\lambda_j(D^*AD)^{1/2} \\leq \\lambda_j(D^*BD)^{1/2}$ for all $j=1,\\dots,n$. Since the matrices are positive semidefinite, the eigenvalues are the singular values and $\\|(D^*AD)^{1/2}\\|_{(k)} \\leq \\|(D^*BD)^{1/2}\\|_{(k)}$ for every Ky-Fan-k norm $k=1,\\dots,n$. By the Fan dominance theorem~\\cite[Theorem IV.2.2]{bhatia1997matrix}, the advertised inequality is true for every unitarily invariant norm.\n \\end{proof}\n}\n\n \n\n\n\\subsection{Proofs of~\\cref{ssec:angles} Theorems}\n\\begin{proof}[\\cref{thm:u}] We tackle each case separately. \\\\\n\\textbf{Bounds for $\\sin\\theta_j$}: The proof is lengthy and proceeds in four steps. We give a great level of detail here, since the proof technique will be applicable to the subsequent proofs.\n\\paragraph{1. Converting an SVD to an EVD} We compute the thin SVD of $(I-\\proj{\\widehat{U}})U_k = K S_U G^*.$ The matrix $$S_U  = \\mathsf{diag}\\,(\\sin\\theta_k,\\dots,\\sin\\theta_1) \\in \\mb{R}^{k\\times k}$$ \ncontains the sine of the canonical angles between the subspaces spanned by the columns of $\\widehat{U}$ and $U_k$~\\cite[Equation (13)]{bjorck1973numerical}. It is readily seen that\n \\begin{equation}\\label{eqn:su} G S_U^2 G^* = U_k^*(I-\\proj{\\widehat{U}})U_k.\\end{equation}\n\n\\paragraph{2. Shrinking space} In~\\cref{alg:randsvd}, we had defined $Y = (AA^*)^qA\\Omega$. It follows that \n\\[ U^*Y = \\bmat{\\Sigma_k^{2q+1} \\\\ & (\\Sigma_\\perp\\Sigma_\\perp^\\top)^q\\Sigma_\\perp} (V^*\\Omega)= \\begin{bmatrix} \\Sigma_k^{2q+1} \\Omega_1 \\\\ (\\Sigma_\\perp\\Sigma_\\perp^\\top)^q\\Sigma_\\perp \\Omega_2\\end{bmatrix},\\]\nwhere from~\\cref{eqn:omegadef}, $\\Omega_1 = V_k^*\\Omega$ and $\\Omega_2 = V_\\perp^*\\Omega$. Next, by~\\cref{ass:main}, $\\Omega_1$ has full row rank and therefore it has a right multiplicative inverse. Define\n\\[ Z \\equiv U^*Y\\Omega_1^\\dagger \\Sigma_k^{-(2q+1)} = \\bmat{I \\\\ F}• \\qquad F \\equiv (\\Sigma_\\perp\\Sigma_\\perp^\\top)^q\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger\\Sigma_k^{-(2q+1)}. \\]\n{Recall that  $Y= QR$ is the thin-QR factorization of $Y$. Let $Q_1R_1$ be the thin-QR factorization of $R\\Omega_1^\\dagger\\Sigma_k^{-(2q+1)}$; here, $Q_1 \\in \\mb{C}^{(k+\\rho) \\times k}, R_1\\in \\mb{C}^{k\\times k}$.  \n\n\nFrom $Q_1Q_1^*\\preceq I$, the conjugation rule implies \n\\[\n\\proj{Z} = U^*QQ_1Q_1^*Q^*U \\preceq U^*QQ^*U = \\proj{U^*Q}.\n\\]\nSince\n$\\mc{R}\\,(U^*Y) = \\mc{R}\\,(U^*Q) = \\mc{R}\\,(U^*\\widehat{U})$, they have the same projectors, so \n\\begin{equation}\\label{eqn:projz}\\proj{Z} \\preceq \\proj{U^*\\widehat{U}} \\qquad I - \\proj{U^*\\widehat{U}} \\preceq I - \\proj{Z}.\\end{equation}\n\n\nPlug in $UU^* = I$ into~\\eqref{eqn:su}, and use~\\eqref{eqn:projz} to obtain\n\\[ U_k^*(I-\\proj{\\widehat{U}})U_k = U_k^*U (I-\\proj{U^*\\widehat{U}}) U^*U_k  \\preceq  \\bmat{ I & 0 }(I-\\proj{Z}) \\bmat{I \\\\ 0}.\\]\n}\n\n\\paragraph{3. Simplifying $\\proj{Z}$} \nSince  $\\proj{Z} = ZZ^\\dagger$, we have\n\\[ \\proj{Z} = \\bmat{I \\\\ F} (I + F^*F)^{-1} \\bmat{I & F^*}, \\]\nfrom which, it can be readily seen that \n\\begin{align} \\nonumber\n \\bmat{ I & 0 }(I-\\proj{Z}) \\bmat{I \\\\ 0} = & \\> I - (I+F^*F)^{-1} \\\\ \\label{eqn:H}\n= &  \\> F^*F (I+F^*F)^{-1}\\equiv H.\n\\end{align}\nNote that $H$ is positive semidefinite. To summarize the story so far, $GS_U^2 G^* \\preceq  H$. \n\\paragraph{4. Applying singular value inequalities}  A straightforward SVD argument shows that the $j$-th singular value of $ H$ satisfies\n\\[  \\sigma_j(H) = \\sigma_j^2(F)/(1+\\sigma_j^2(F)) \\quad j=1,\\dots,k. \\] \nThe singular value inequalities~\\cref{eqn:singprod} imply \n\\[ \\sigma_j(F) \\leq \\sigma_1( \\Sigma_\\perp\\Sigma_\\perp^\\top)^q\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger)\\sigma_j(\\Sigma_k^{-2q-1}) \\leq \\left(\\frac{\\sigma_{k+1}}{\\sigma_{k-j+1}}\\right)^{2q+1} \\normtwo{\\Omega_2\\Omega_1^\\dagger} .\\]\nPlugging this inequality into $\\sigma_j(H)$\n\\[ \\sigma_j^2(H) \\> \\leq  \\>  \\frac{\\gamma_{k-j+1}^{4q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2}{1 + \\gamma_{k-j+1}^{4q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2}\\, \\qquad j=1,\\dots,k. \\]\nSince $GS_U^2 G^* \\preceq  H$, Weyl's theorem implies $\\sin^2 \\theta_{k-j+1} \\leq \\sigma_j^2(H)$. \nTake square roots on both sides and rename $j \\leftarrow k-j+1$ to get the desired result.\n\n\n\n\\textbf{Bounds for $\\sin\\nu_j$}:  Let $GS_V^2G^*$ be the eigenvalue decomposition of $V_k^*(I-\\proj{\\widehat{V}})V_k$. Note that the diagonals of $S_V$ are the sine of the canonical angles $\\angle(V_k,\\widehat{V})$. Since $\\widehat{V}$\nis obtained from the thin SVD of  $A^*Q$, $\\mc{R}\\,(A^*Q) =\n\\mc{R}\\,(\\widehat{V})$ and $\\proj{\\widehat{V}} = \\proj{A^*Q}$, since an orthogonal projection\nmatrix is uniquely determined by the range. Next, consider $\\widehat{Z}$ defined as \n\\begin{equation}\\widehat{Z} \\equiv \\Sigma^\\top U^*Y\\Omega_1^\\dagger \\Sigma_k^{-2q-2} = \\begin{bmatrix}I \\\\ \\widehat{F}\\end{bmatrix}• \\qquad \\widehat{F} {\\equiv} (\\Sigma_\\perp^\\top\\Sigma_\\perp)^{q+1}\\Omega_2\\Omega_1^\\dagger\\Sigma_k^{-2q-2}, \\label{zhat}\n\\end{equation}• \nfrom $(AV)^*Q= \\Sigma^*U^*Q $, it can be verified that \n\\[ \\mc{R}\\,(\\widehat{Z} ) \\subset \\mc{R}\\,(\\Sigma^\\top U^*Y) = \\mc{R}\\,(\\Sigma^\\top U^*Q) = \\mc{R}\\,((AV)^*Q). \\]\nUsing an argument similar to~\\cref{eqn:projz}, we obtain\n\\[ V_k^*V(I-\\proj{\\widehat{V}})V^* V_k   \\preceq V_k^*V(I-\\proj{\\widehat{Z}})V^* V_k = \\bmat{I & 0} (I-\\proj{\\widehat{Z}}) \\bmat{I \\\\0}. \\] \nThe right hand side simplifies to $I - (I +\\widehat{F}^*\\widehat{F})^{-1}$. The rest of the proof is similar to that of the proof for $\\sin\\theta_j$.\n\\end{proof}\n\n\n\n\n\n\\begin{proof}[\\cref{thm:theta2F}]\nWith the notation of \\cref{thm:u}, we follow steps 1-3 of the proof to obtain $$GS_U^2 G^* \\preceq  H \\preceq F^*F.$$\nSince the square root preserves partial ordering, implies $GS_U G^* \\preceq   (F^*F)^{1/2}$. {Note that $(F^*F)^{1/2}$ and $F$ have the same nonzero singular values. Therefore, $$\\uninorm{\\sin\\angle(U_k,\\widehat{U})} \\leq\\uninorm{(F^*F)^{1/2}} = \\uninorm{F}.$$ \nBy using strong sub-multiplicativity of the unitarily invariant norm, we have\n$$\\uninorm{\\sin\\angle(U_k,\\widehat{U})} \\leq \\gamma_k^{2q}\\normtwo{\\Omega_2\\Omega_1^\\dagger} \\frac{\\uninorm{\\Sigma_\\perp}}{\\sigma_k}.$$\n} \n\\end{proof}\n\n\n\n\n\\subsection{Proofs of~\\cref{ssec:sintheta} Theorems}\n\n\n\\begin{proof}[\\cref{thm:gensintheta}]\nLet $X = (I-\\proj{\\widehat{U}_k})\\proj{U_k}$ and $Y = (I-\\proj{\\widehat{V}_k})\\proj{V_k}$. In decreasing order, the singular values of $X$ and $Y$ are $\\{ \\sin \\theta_j'\\}_{j=1}^k$ and $ \\{\\sin \\nu_j'\\}_{j=1}^k$ respectively. Let $B \\equiv \\widehat{A} - \\widehat{A}_k$. First, we observe that\n\\begin{align*}\nE_{12} = & \\> (I-\\proj{\\widehat{U}_k})(A - \\widehat{A})\\proj{V_k} \\\\ \n= & \\> (I-\\proj{\\widehat{U}_k})A_k - (I-\\proj{\\widehat{U}_k}) \\widehat{A} \\proj{V_k} \\\\\n= &\\> (I-\\proj{\\widehat{U}_k})\\proj{U_k}A_k - (\\widehat{A} - \\widehat{A}_k)\\proj{V_k} \\\\\n=& \\> XA_k - B(I-\\proj{\\widehat{V}_k})\\proj{V_k} =XA_k - BY.\n\\end{align*}•\nA similar calculation shows that $E_{21} = X^*B - A_kY^*$. From the first relation, since $\\mathsf{rank}\\,(A) \\geq k$, we have \n\\[ XA_k A_k^\\dagger = (E_{12} + BY) A_k^\\dagger.\\]\nBut $A_k A_k^\\dagger = \\proj{U_k}$ and $X\\proj{U_k} = X$. Applying~\\cref{eqn:singprod}, we have  \n\\[ \\sigma_j(X) \\leq (\\|E_{12}\\|_2  + \\|B\\|_2\\|Y\\|_2)/\\sigma_{k-j+1}(A) \\qquad j=1,\\dots,k.\\]\nA similar argument gives\n\\[\\sigma_j(Y) \\leq (\\normtwo{E_{21}}  + \\normtwo{B}\\normtwo{X})/\\sigma_{k-j+1}(A) \\qquad j=1,\\dots,k. \\]\nCombining these relations \n\\[ \\max \\{ \\sigma_j(X), \\sigma_j(Y)\\} \\leq \\frac{\\max\\{ \\|E_{21}\\|_2, \\|E_{12}\\|_2\\}}{\\sigma_{k-j+1}(A)}  + \\frac{\\|B\\|_2}{\\sigma_{k-j+1}(A)} \\max\\{\\|X\\|_2,\\|Y\\|_2\\}. \\]\nRecognize that $\\|B\\|_2 = \\sigma_{k+1}(\\widehat{A})$. \nApplying~\\cref{eqn:sintheta} in the spectral norm simplifies the expression since \n\\[ \\frac{1}{\\sigma_{k-j+1}(A)}\\left(1  + \\frac{\\sigma_{k+1}(\\widehat{A})}{\\sigma_{k}(A) - \\sigma_{k+1}(\\widehat{A})}\\right) = \\frac{\\sigma_k(A)}{\\sigma_{k-j+1}(A)(\\sigma_{k}(A) - \\sigma_{k+1}(\\widehat{A}))}. \\]\nTherefore, \n\\[\\max \\{ \\sigma_j(X), \\sigma_j(Y)\\} \\leq \\frac{\\sigma_k(A)}{\\sigma_{k-j+1}(A)}\\frac{\\max\\{ \\normtwo{E_{21}}, \\normtwo{E_{12}}\\}}{\\zeta}.\\]\nNow $\\sigma_j(X) = \\sin\\theta_{k-j+1}'$ and $\\sigma_j(Y) = \\sin\\nu_{k-j+1}'$. Rename $j\\leftarrow k-j+1$ to finish. \n\\end{proof}\n\n\n\\begin{proof}[\\cref{thm:sintheta}] We tackle each case independently. \\\\\n\\textbf{Unitarily invariant norms}: Our proof involves simplifying each term in~\\cref{eqn:sintheta}, and~\\cref{eqn:e1} and has several steps.\n\n\\paragraph{1. Simplifying the gap} Recall $\\zeta = \\sigma_k(A) - \\sigma_{k+1}(\\widehat{A})$ and $\\widehat{A} = QQ^*A$. From the first part of~\\cref{thm:sigma} \n\\[ \\zeta = \\sigma_k(A) - \\sigma_{k+1}(\\widehat{A}) \\geq \\sigma_k(A) - \\sigma_{k+1}(A).\\]\n\t\\paragraph{2. Simplifying $\\uninorm{E_{12} }$} First observe that $A\\proj{V_k} = A_k$. So $$E_{12} =(I-\\proj{\\widehat{U}_k})(I-QQ^*)A\\proj{V_k} =  (I-\\proj{\\widehat{U}_k})(I-QQ^*)A_k.$$ Then applying~\\cref{eqn:nnorm_rankr} along with sub-multiplicativity gives\n\\[ \\uninorm{E_{12} } \\leq \\uninorm{(I-QQ^*)A_k}  \\leq \\gamma_k^{2q}\\uninorm{\\Sigma_\\perp\\Omega_2\\Omega_1^\\dagger } \\leq \\gamma_k^{2q} \\uninorm{\\Sigma_\\perp}  \\normtwo{\\Omega_2\\Omega_1^\\dagger }. \\]\n\\paragraph{3. Simplifying $\\uninorm{E_{21} }$} First, $E_{21} = \\proj{U_k}(I-QQ^*)A\\proj{\\widehat{V}_k}$, and since $\\normtwo{\\proj{U_k}(I-QQ^*)} = \\normtwo{\\sin\\angle(U_k,\\widehat{U})}$, \n\\[ \\uninorm{E_{21} } \\leq \\normtwo{\\sin\\angle(U_k,\\widehat{U})} \\uninorm{ (I-QQ^*)A},\\]\nbecause of strong sub-multiplicativity. Applying \\cref{thm:u} and~\\cref{eqn:nnorm_rankr}\n\\[  \\uninorm{E_{21}} \\leq \\frac{\\gamma_k^{2q+1} \\normtwo{\\Omega_2\\Omega_1^\\dagger}}{\\sqrt{1+\\gamma_k^{4q+2} \\| \\Omega_2\\Omega_1^\\dagger\\|_2^2}} \\left(1  + \\gamma_k^{2q}\\normtwo{\\Omega_2\\Omega_1^\\dagger} \\right)\\uninorm{\\Sigma_\\perp}.\\]\nLet $\\beta = \\gamma_k^{2q}\\normtwo{\\Omega_2\\Omega_1^\\dagger}$. Then for $\\beta \\geq 0$, since $\\gamma_k < 1$\n\\[ \\frac{\\gamma_k(1+\\beta)}{\\sqrt{1+\\gamma_k^2\\beta^2}} \\leq \\frac{1+\\gamma_k\\beta}{\\sqrt{1+\\gamma_k^2\\beta^2}} \\leq \\sqrt{2}. \\] \nTherefore, $\\uninorm{E_{21}} \\leq \\sqrt{2}\\gamma_k^{2q} \\uninorm{\\Sigma_\\perp} \\|\\Omega_2\\Omega_1^\\dagger\\|_2  $.\n\\paragraph{4. Putting everything together} Plugging in the intermediate quantities into~\\cref{eqn:sintheta}, we have\n \\[ \\max\\left\\{ \\uninorm{ \\sin \\angle(U_k,\\widehat{U}_k)} ,\\uninorm{\\sin\\angle(V_k,\\widehat{V}_k)} \\right\\} \\leq \\sqrt{2}\\gamma_k^{2q} \\normtwo{\\Omega_2\\Omega_1^\\dagger} \\frac{ \\uninorm{\\Sigma_\\perp}}{\\sigma_{k}-\\sigma_{k+1}}. \\]\nDividing the numerator and denominator by $\\sigma_k$ proves the stated result for unitarily invariant norms. \n\n\n\\textbf{Spectral/Frobenius norms}: Let $\\norm{\\cdot}{\\xi}$ denote the spectral and Frobenius norms. The first two steps are identical to the proof for unitarily invariant norms. For the third step, using~\\cref{eqn:nnorm_rankr}\n\\[ \\norm{E_{21}}{\\xi} \\leq \\frac{\\gamma_k^{2q+1} \\| \\Omega_2\\Omega_1^\\dagger\\|_2}{\\sqrt{1+\\gamma_k^{4q+2} \\| \\Omega_2\\Omega_1^\\dagger\\|_2^2}} \\norm{\\Sigma_\\perp}{\\xi} \\sqrt{1+\\gamma_k^{4q} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2}.\\]\nWith $\\beta$ defined as before, since $\\gamma_k < 1$,  ${\\sqrt{\\gamma_k^2+\\gamma_k^2\\beta^2}}/{\\sqrt{1+\\gamma_k^2\\beta^2}} \\leq  1$.\nTherefore, $$\\norm{E_{21}}{\\xi} \\leq \\gamma_k^{2q} \\norm{\\Sigma_\\perp}{\\xi}\\normtwo{\\Omega_2\\Omega_1^\\dagger}.$$ The rest of the proof is the same. \n\n\n\\textbf{Canonical angles}: The proof combines~\\cref{thm:gensintheta} with the above analysis for the spectral norm. The right hand side contains the term \n$$\\max\\left\\{ \\normtwo{ \\sin \\angle(U_k,\\widehat{U}_k)} ,\\normtwo{\\sin\\angle(V_k,\\widehat{V}_k)} \\right\\}.$$ \nThe rest of the proof involves some simple manipulations.\n\\end{proof}\n\n\n\n\\begin{proof}[\\cref{t_singleuv}]\nWe first address~\\cref{eqn:singleuv}. Following the steps of the proof of~\\cref{thm:u}, we have \n\\[ \\sin^2 \\angle (u_j,\\widehat{U}) = u_j^*U(I-\\proj{U^*Q})U^*u_j \\preceq \\bmat{e_j^\\top & 0 } (I - \\proj{Z}) \\bmat{e_j \\\\ 0}, \\]\nwhere $e_j$ is the $j$--th  column of the $k\\times k$ identity matrix. Therefore, we have $\\sin^2 \\angle (u_j,\\widehat{U}) \\leq e_j^\\top He_j$,\nwhere $H$ was defined in~\\cref{eqn:H}. The inequality $H \\preceq F^*F$ implies\n\\begin{align*}\n  \\sin^2 \\angle (u_j,\\widehat{U}) \n \\leq& \\>\\sigma_{j}^{-4q-2}  \\normtwo{(\\Sigma_\\perp\\Sigma_\\perp^\\top)^q\\Sigma_\\perp(\\Omega_2\\Omega_1^\\dagger)e_j}^2 \\\\\n \\leq & \\>  \\gamma_{j}^{4q+2}\\normtwo{\\Omega_2\\Omega_1^\\dagger}^2. \n\\end{align*}\nTaking square-roots on both sides gives the desired results. The strategy for bounding the canonical angles $\\sin\\angle (v_j,\\widehat{V})$ is very similar and will be omitted.\n\nWe now address~\\cref{eqn:svs}, which is a straightforward application of~\\cite[Theorem 2.5]{hochstenbach2004harmonic}. Let $\\proj{\\mc{U}} = QQ^*$ and $\\proj{\\mc{V}} = I$. Then, in our notation, this result takes the form\n\\[ \\max\\left\\{\\sin\\angle(u_j,\\widehat{u}_j),\\sin\\angle(v_j,\\widehat{v}_j)\\right\\} \\leq \\sqrt{1 + 2\\frac{\\tilde\\gamma'^2}{\\tilde\\delta^2}}\\max\\left\\{\\sin\\angle(u_j,\\widehat{U}),\\sin\\angle(v_j,I)\\right\\}. \\]\nwhere $\\tilde\\gamma' = \\max\\{0, \\normtwo{(I-QQ^*)A}\\}$ and $\\tilde\\delta$ is as defined in the statement of the theorem. \\cref{thm:lowrank} for the spectral norm implies $\\tilde\\gamma'\\leq \\tilde\\gamma$, whereas \\cref{t_singleuv} implies \n\\[ \\max\\left\\{\\sin\\angle(u_j,\\widehat{U}),\\sin\\angle(v_j,I)\\right\\} \\leq \\gamma_j^{2q}\\normtwo{\\Omega_2\\Omega_1^\\dagger}.\\]\nPlug in the intermediate steps to obtain the desired bound.  \n\\end{proof}\n\n\n\\begin{proof}[\\cref{thm:can_gauss}]\nIn~\\cref{thm:u}, bounds for $\\normtwo{\\Omega_2\\Omega_1^\\dagger}$ are available in the literature. From the proof of~\\cite[Theorem 10.6]{HMT09} we find the inequality\n$$\\mb{E}\\,\\normtwo{\\Omega_2\\Omega_1^\\dagger} \\leq C_e,$$ where the constant $C_e$ was defined in~\\cref{eqn:ce}. Let $\\alpha > 0$ be a constant. The map $x \\mapsto x/\\sqrt{1+\\alpha x^2}$ is convex. Therefore, by Jensen's inequality the results in expectation follow. \n\nFor the concentration inequalities,~\\cite[Theorem 5.8]{gu2015subspace} showed that $\\normtwo{\\Omega_2\\Omega_1^\\dagger} \\leq C_d$ with a probability at least $1-\\delta$. Here, $C_d$ was defined in~\\cref{eqn:cd}. Plug into~\\cref{thm:u} to obtain the desired bounds.\n\\end{proof}\n\n\n\n\n\\subsection{Proofs of~\\cref{sec:aux} Theorems}\n \n \n\\begin{proof}[\\cref{thm:lowrank_schatten}]  \n\n\\textbf{Proof of~\\cref{eqn:nnorm}}:\nUsing the unitary invariance of the norms\n\\[ \\uninorm{(I-\\proj{Q})A} = \\uninorm{(I-\\proj{U^*Q})\\Sigma}  = \\uninorm{(\\Sigma^\\top(I-\\proj{U^*Q})\\Sigma)^{1/2}}.\\]\n We  use~\\cref{eqn:projz} combined with \\cref{lemma:schatten} to obtain \n $$\\uninorm{(\\Sigma^\\top(I-\\proj{U^*Q})\\Sigma)^{1/2}} \\leq \\uninorm{(\\Sigma^\\top(I-\\proj{Z})\\Sigma)^{1/2}}.$$\nWith $M_1 \\equiv I-(I+F^*F)^{-1}$ and $M_2 \\equiv I - F(I+F^*F)^{-1}F^*$, then $\\Sigma^\\top(I-\\proj{Z})\\Sigma$ simplifies as\n\\begin{equation}\\label{eqn:block} \\Sigma^\\top (I-\\proj{Z})\\Sigma = \\bmat{\\Sigma_k M_1\\Sigma_k & * \\\\ * &  \\Sigma_\\perp^\\top M_2\\Sigma_\\perp}.\\end{equation}\nThe square root function is  concave on $[0,\\infty)$ and $\\Sigma^\\top(I-\\proj{Z})\\Sigma$ is positive semidefinite. Therefore, an extension to Rotfel'd's theorem says~\\cite[Theorem 2.1]{lee2011extension} \n\\[ \\uninorm{(\\Sigma^\\top(I-\\proj{Z})\\Sigma)^{1/2}} \\leq \\uninorm{(\\Sigma_k^\\top M_1\\Sigma_k)^{1/2}} +  \\uninorm{(\\Sigma_\\perp^\\top M_2\\Sigma_\\perp)^{1/2}}. \\] \n{\nUse the inequalities $M_1 \\preceq F^*F$ and $M_2\\preceq I$, along with \\cref{lemma:schatten} gives \n\\begin{equation}\n\\begin{split}\n \\uninorm{(I-\\proj{Q})A} \\leq & \\>  \\uninorm{(\\Sigma_kF^*F\\Sigma_k)^{1/2}} + \\uninorm{(\\Sigma_\\perp\\Sigma_\\perp)^{1/2}}\\\\  \n\\leq & \\> \\uninorm{F\\Sigma_k} + \\uninorm{\\Sigma_\\perp}.\n\\end{split}\n\\end{equation}\nUse $F\\Sigma_k = (\\Sigma_\\perp\\Sigma_\\perp)^q\\Sigma_\\perp \\Omega_2\\Omega_1^\\dagger \\Sigma_{k}^{-2q}$ and the sub-multiplicativity to obtain the advertised bounds.}\n\n\n\n\\textbf{Proof of~\\cref{eqn:nnorm_rankr}}: The proof for~\\cref{eqn:nnorm_rankr} is similar and is omitted. The main observation is that $A_k$ has only $k$\nnonzero singular values. \n\n\n\\textbf{Proof of~\\cref{eqn:nnorm_rankr2}}: We follow the strategy in~\\cite[Section 3.3]{drineas2018low}. {Recall that $B_k$ is the best rank-$k$ approximation to $B = Q^*A$. With the notation in~\\cref{alg:truncation}, note that \n$$QB_k = Q\\widehat{U}_{B,k} \\widehat{U}_{B,k}^*B = \\widehat{U}_k \\widehat{U}_{k}^*A = \\proj{\\widehat{U}_k}A,$$ the triangle inequality gives   \n\\[ \\uninorm{(I-\\proj{\\widehat{U}_k})A} \\leq \\uninorm{(I-\\proj{\\widehat{U}_k})A_k} + \\uninorm{(I-\\proj{\\widehat{U}_k})A_{\\perp}}.  \\]\n}\nSince $A_k = \\proj{U_k}A_k$, applying strong sub-multiplicativity \n\\[ \\uninorm{(I-\\proj{\\widehat{U}_k})A} \\leq \\uninorm{(I-\\proj{\\widehat{U}_k})\\proj{U_k}} \\normtwo{A_k} + \\uninorm{A_{\\perp}}.\\] \nWe recognize that $\\uninorm{(I-\\proj{\\widehat{U}_k})\\proj{U_k}} = \\uninorm{\\sin\\angle(U_k,\\widehat{U}_k)}$, apply~\\cref{thm:sintheta} to complete the proof. \n\\end{proof}\n\\begin{proof}[\\cref{thm:lowrank}]\nThe proof is similar to that of the proof of~\\cref{thm:lowrank_schatten}. Consider the term of interest $\\schattenp{ (I-QQ^*)A}^2$, which can be simplified to \n\\[ \\schattenp{ (I-QQ^*)A }^2 = \\uninorm{A^*(I-QQ^*) A}_{p/2} = \\uninorm{ \\Sigma^\\top (I - \\proj{U^*Q}) \\Sigma}_{p/2}.\\] \nThe first equality holds only for $p\\geq 2$, whereas the last equality follows because of the unitary invariance.  As in the proof of~\\cref{thm:u}, we have \n \\[ Z = U^*Y\\Omega_1^\\dagger \\Sigma_k^{-(2q+1)}\\qquad F = (\\Sigma_\\perp \\Sigma_\\perp^\\top)^q \\Sigma_\\perp\\Omega_2 \\Omega_1^\\dagger \\Sigma_k^{-(2q+1)}.\\]\nThe use of~\\cref{eqn:projz} and~\\cref{lemma:schatten} ensures \n \\[\\uninorm{ \\Sigma^\\top (I - \\proj{U^*Q}) \\Sigma}_{p/2} \\leq \\uninorm{\\Sigma^\\top (I - \\proj{Z}) \\Sigma}_{p/2}.\\]\n {We apply~\\cite[Theorem 2.1]{lee2011extension} to~\\cref{eqn:block} with $f(t) =t$ to obtain\n \\begin{equation*}\n\\begin{aligned}\\uninorm{\\Sigma^\\top (I - \\proj{Z}) \\Sigma}_{p/2} \\leq & \\> \\uninorm{\\Sigma_k M_1 \\Sigma_k}_{p/2} +\\uninorm{\\Sigma_\\perp^\\top M_2  \\Sigma_\\perp}_{p/2}\\\\\n\\leq & \\>\\uninorm{\\Sigma_k F^*F \\Sigma_k}_{p/2} + \\uninorm{\\Sigma_\\perp^\\top  \\Sigma_\\perp}_{p/2}\\\\\n= & \\> \\schattenp{F\\Sigma_k}^2 + \\schattenp{\\Sigma_\\perp}^2. \\end{aligned} \\end{equation*}\nWe have used $M_1 \\preceq F^*F$ and $M_2 \\preceq I$.} The rest of the proof is similar to that of~\\cref{thm:lowrank_schatten}. \n\\end{proof}\n\n\n\n\n\\begin{proof}[\\cref{thm:sigma}]\n{The proof makes heavy use of the partial ordering which was reviewed in the start of \\cref{sec:proofs}.} From the inequality $I \\succeq QQ^*$, the conjugation rule gives $$A^*A \\succeq A^*QQ^*A.$$ Then, Weyl's theorem implies $\\lambda_j(A^*A) \\geq \\lambda_j(A^*QQ^*A)$ for $j=1,\\dots,k$. Relating the eigenvalues to the singular values proves the first inequality.\n\nFor the second inequality consider again $A^*QQ^*A$. With the aid of~\\cref{eqn:projz} \n\\begin{equation}\\label{eqn:inters} A^*QQ^*A = V\\Sigma^\\top\\proj{U^*Q}\\Sigma V^* \\succeq V\\Sigma^\\top\\proj{Z}\\Sigma V^*. \\end{equation}\nTherefore, $\\lambda_j(A^*QQ^*A) \\geq \\lambda_j(V\\Sigma^\\top\\proj{Z}\\Sigma V^*)$ for $j =1,\\dots,k$. Since $V\\Sigma^\\top\\proj{Z}\\Sigma V^*$ and $\\Sigma^\\top\\proj{Z}\\Sigma$ are similar, they share the same eigenvalues. It can be readily shown that \n\\[ \\Sigma^\\top\\proj{Z}\\Sigma = \\bmat{ \\Sigma_k (I+F^*F)^{-1}\\Sigma_k & * \\\\ * & *} .\\]\nFor $j=1,\\dots,k$, the eigenvalues of $A^*Q^*QA$ satisfy \n\\begin{equation}\\label{eqn:ordering} \\lambda_j(A^*QQ^*A)  \\geq  \\lambda_j(V\\Sigma^\\top\\proj{Z}\\Sigma V^*) \\geq  \\lambda_j(\\Sigma_k (I+F^*F)^{-1}\\Sigma_k).\n\\end{equation}\n{The second inequality follows from the Cauchy interlacing theorem~\\cite[Section 10-1]{Par80}.  Applying the properties of partial ordering, we obtain\n\\[ F^*F \\preceq \\sigma_{k+1}^{4q+2} \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2 \\Sigma_k^{-(4q+2)} = \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2 \\Gamma_k^{4q+2}, \\]\nwhere $\\Gamma_k = \\mathsf{diag}\\,(\\gamma_1,\\dots,\\gamma_k)$ is a diagonal matrix with the singular value gaps. Furthermore, \n\\[ \\Sigma_k (I+F^*F)^{-1}\\Sigma_k \\succeq \\Sigma_k(I + \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2\\Gamma_k^{4q+2})^{-1}\\Sigma_k.\\]\nSince the diagonal matrix on the right hand side has its singular values on the diagonals; this fact, combined with~\\cref{eqn:ordering} gives for $j=1,\\dots,k$\n \\[  \\sigma_j^2(Q^*A) = \\lambda_j(A^*QQ^*A)  \\geq  \\lambda_j(\\Sigma_k (I+F^*F)^{-1}\\Sigma_k) \\geq \\frac{\\sigma_j^2(A)}{1 + \\normtwo{\\Omega_2\\Omega_1^\\dagger}^2 \\gamma_{j}^{4q+2}}.\\]\n Taking square-roots, we obtain the desired result.\n}\n\\end{proof}\n\n\n\\section{Numerical Results}\n\n\n\\subsection{Test matrices}\\label{ssec:test} To demonstrate the performance of the bounds, we use the following test matrices\n\\begin{enumerate}\n\\item \\textbf{Controlled gap} The first set of test matrices $A \\in \\mathbb{R}^{3000\\times 300}$ are constructed using the formula\n\\[ A = \\sum_{j=1}^{r}\\frac{\\text{gap}}{j} \\, x_j y_j^\\top +\n  \\sum_{j=r+1}^{300}\\frac{1}{j} \\,x_j y_j^\\top,\\]\nwhere $x_j \\in \\mathbb{R}^{3000}$ and $y_j \\in \\mathbb{R}^{300}$ are sparse random vectors with non-negative entries generated using the MATLAB commands \\verb|sprand(3000,1,0.025)| and \\verb|sprand(300,1,0.025)| respectively. The formula above is not an SVD, since the vectors do not form an orthonormal set. Nonetheless, the singular values decay like $1/j$ and the gap between the singular values between $15$ and $16$ is controlled by the parameter $\\text{gap}$. We consider three cases:  \n\\begin{enumerate}\n\\item Small gap (GapSmall) $\\text{gap} = 1$,\n\\item Medium gap (GapMedium) $\\text{gap} = 2$,\n\\item Large gap (GapLarge) $\\text{gap} = 10$.\n\\end{enumerate}•\n\n\\begin{figure}[!ht]\\centering\n\\includegraphics[scale=0.25]{figs/sv_gap}\n\\includegraphics[scale=0.25]{figs/sv_noise}\n\\includegraphics[scale=0.25]{figs/sv_decay}\n\\caption{Singular value of the matrices from the (left) `Controlled Gap' example, (right) `Low-rank plus noise' example, (below) `Low-rank plus decay' example.  }\n\\label{fig:svs}\n\\end{figure}•\n\n\\item \\textbf{Low-rank plus noise} The matrices are of the form\n\\[ A = \\begin{bmatrix} I_r & 0 \\\\ 0 & 0\\end{bmatrix} + \\sqrt{\\frac{\\gamma_n r}{2n^2}}(G + G^\\top),\\]\nwhere $G \\in \\mathbb{R}^{n\\times n}$ is a random Gaussian matrix. \nWe consider three cases:  \n\\begin{enumerate}\n\\item Small noise (NoiseSmall) $\\gamma_n = 10^{-2}$,\n\\item Medium noise (NoiseMedium) $\\gamma_n = 10^{-1}$,\n\\item Large noise (NoiseLarge) $\\gamma_n = 1$.\n\\end{enumerate}•\n\\item \\textbf{Low-rank plus decay} The matrices take the form\n\\[ A = U\\mathsf{diag}\\,(\\underbrace{1,1,\\dots,1}_r,2^{-d},3^{-d}\\dots,(n-r+1)^{-d})V^*.\\]\nThe unitary matrices $U,V$ are obtained by drawing a random Gaussian matrix, and taking its QR factorization. We distinguish between the following cases\n\\begin{enumerate}\n\\item Slow decay (DecaySlow): $d = 0.5$,\n\\item Medium decay (DecayMedium): $d = 1.0$,\n\\item Fast decay  (DecayFast): $d=2.0$.\n\\end{enumerate}\n\\end{enumerate}\nThe first example is adapted from~\\cite{sorensen2016deim}, whereas the second and third examples are drawn from~\\cite{tropp2017practical}. In all the examples, the random matrices were fixed by setting the random seed and we the set the parameter $r=15$. The singular values of all the test matrices are plotted in~\\cref{fig:svs}. \n\\begin{figure}[!ht]%\n\\centering\n\\subfigure[GapSmall]{%\n\\label{fig:cana1}%\n\\includegraphics[scale=0.24]{figs/canon_gap_1}}%\n\\subfigure[GapMedium]{%\n\\label{fig:canb1}%\n\\includegraphics[scale=0.24]{figs/canon_gap_2}}%\n\\subfigure[GapLarge]{%\n\\label{fig:canc1}%\n\\includegraphics[scale=0.24]{figs/canon_gap_3}}%\n\\\\\n\\subfigure[NoiseSmall]{%\n\\label{fig:cana2}%\n\\includegraphics[scale=0.24]{figs/canon_noise_1}}%\n\\subfigure[NoiseMedium]{%\n\\label{fig:canb2}%\n\\includegraphics[scale=0.24]{figs/canon_noise_2}}%\n\\subfigure[NoiseLarge]{%\n\\label{fig:canc2}%\n\\includegraphics[scale=0.24]{figs/canon_noise_3}}%\n\\\\\n\\subfigure[DecaySlow]{%\n\\label{fig:cana3}%\n\\includegraphics[scale=0.24]{figs/canon_decay_1}}%\n\\subfigure[DecayMedium]{%\n\\label{fig:canb3}%\n\\includegraphics[scale=0.24]{figs/canon_decay_2}}%\n\\subfigure[DecayFast]{%\n\\label{fig:canc3}%\n\\includegraphics[scale=0.24]{figs/canon_decay_3}}%\n\\caption{Plots of $\\sin \\theta_j $ for $j=1,\\dots,k$. The test matrices were described in \\cref{ssec:test}. The target rank $k=25$ and an oversampling parameter of $20$ was chosen for all the experiments. The solid lines correspond to the computed values, the dashed lines correspond to bounds obtained using~\\cref{thm:u}. The parameter $q$ corresponds to the number of subspace iterations. }\n\\label{fig:can}\n\\end{figure}\n\\subsection{Canonical angles}\nFor the first numerical example, we use the $9$ test matrices in~\\cref{ssec:test}. For each matrix, we chose an oversampling parameter $\\rho=20$ and the target rank $k$ was chosen to be $25$. The starting guess $\\Omega$ was taken to be a random Gaussian matrix.  \n\n\n\\subsubsection{No extraction} We plot the canonical angles $\\sin\\angle(U_k,\\widehat{U})$ in solid lines, the corresponding bounds from~\\cref{thm:u} are also plotted in dashed lines.  The results are displayed in~\\cref{fig:can}. We make the following general observations:\n{\n\\begin{itemize}\n\\item The influence of the subspace iterations on the canonical angles is clear: the angles become smaller as the number of iterations $q$ increases. This implies that the subspace is becoming more accurate. \n\\item If there is a large singular value gap in the spectrum, this means that all the canonical angles below that index are captured accurately. This is prominently seen in~\\cref{fig:canc1}, in which there is a large gap between singular values $15$ and $16$. Similar observations can be made in the other figures. \n\\item As the decay rate of the singular values increases, the corresponding canonical angles become smaller. \n\\item  In most figures the bounds are qualitatively informative, but in some figures, the bounds are also quantitatively accurate (e.g., GapLarge).\n\\item  Similar results were observed for $\\sin\\angle(V_k,\\widehat{V})$ and, therefore, omitted. \n\\end{itemize}•\n}\n{\nWe now make observations specific to the test examples: \n\\begin{description}\n\\item [1. Gap examples] The computed canonical angles decrease as the gap increases, and with more iterations. The test matrices (GapMedium and GapLarge) have both a decay in the singular values and a prominent singular value gap between indices $15$ and $16$.   These matrices satisfy the assumptions of our analysis, and therefore the bounds can be expected to be good. We see that as the size of the gap increases, the bounds become more accurate in accordance with Theorem~\\ref{thm:u}. GapSmall has decay in the singular values but no special singular value gap. Even in this case, the bounds are qualitatively good.\n\\item [2. Noise examples]  NoiseSmall is close to a low-rank matrix and  there is a large singular value gap at index $15$. For this example, the bounds are qualitatively good. As the level of noise increases, the gap decreases and therefore, the computed angles increase, as predicted by~\\cref{thm:u}. The bounds are uninformative for $q=0$, but qualitatively good for $q=1$ and $2$. Compared to the Gap examples, the bounds are not as sharp since there is very little decay in the singular values. \n\\item [3. Decay examples] In these examples, the singular values decay beyond index $15$ but there is no prominent gap. As the rate of decay increases, in general, the canonical angles decrease. It is also seen that the bounds are qualitatively accurate (except for $q=0$). \n\\end{description}\n }\n\n\n\\subsubsection{Extraction step} Our next experiment tests the effect of the extraction step on the accuracy of the canonical angles. We now compute $\\sin\\theta_j'$ and $\\sin\\nu_j'$ for the test matrices described in~\\cref{ssec:test}. We plot the quantities $\\max\\{\\sin\\theta_j',\\sin\\nu_j'\\}$ for $j=1,\\dots,k$ in solid lines. The corresponding bounds from~\\cref{thm:sintheta} are plotted in dashed lines. Here, the target rank was chosen to be $k=15$, to exploit the singular value gap in the matrices. We make the following general observations:\n{\n\\begin{itemize}\t\n\\item The extraction step did not significantly affect the canonical angles and the accuracy is comparable to~\\cref{fig:can}.  The subspaces are more accurate as the number of iterations increase, and if there is a large singular value gap at index $j$, then the canonical angles with index $j' < j$  are captured accurately.\n\\item Although the canonical angles are small, compared to~\\cref{thm:u}, the bounds in~\\cref{thm:sintheta} are not as accurate. One reason is that the upper bounds in~\\cref{thm:u} are at most $1$, but the bounds in~\\cref{thm:sintheta} are allowed to be greater than $1$. Furthermore, the bound in~\\cref{thm:sintheta} has the factor $1/(1-\\gamma_k)$ in the denominator, which can be quite large when there is a small singular value gap. It may be possible to derive better bounds, but we could not immediately see how to derive them. \n\\item We also compared the accuracy of the individual singular vectors (not shown here). The results and the conclusions are similar. \n\\end{itemize}\n}\n{\nWe now make observations specific to the test examples: \n\\begin{description}\n\\item [1. Gap examples] The behavior of the computed canonical angles is very similar to that without the extraction step. In general, the angles decrease as the gap increases. When the parameter $\\text{gap}$ is small, the singular value ratio $\\gamma_k$ is large, and $(1-\\gamma_k)^{-1}$ is small. This explains why the bounds are bad for GapSmall and GapMedium, and show little improvement with more subspace iterations. Only for the GapLarge example with $q=0$, the bounds are qualitatively good. \n\\item [2. Noise examples] The computed canonical angles decrease as the noise decreases. In all three examples, the  bounds are qualitatively good. The bounds are better for NoiseSmall and NoiseMedium because the singular value gap between indices $15$ and $16$ is bigger than that for NoiseLarge. \n\\item [3. Decay examples] The computed canonical angles become smaller as the decay of the singular values increases. In these examples, there is no prominent gap, so the bounds don't capture the behavior well. However, the computed angles are small, and the subspace is accurate. \n\\end{description}\n }\n\n\n\\begin{figure}[!ht]%\n\\centering\n\\subfigure[GapSmall]{%\n\\label{fig:exta1}%\n\\includegraphics[scale=0.24]{figs/extract_gap_1}}%\n\\subfigure[GapMedium]{%\n\\label{fig:extb1}%\n\\includegraphics[scale=0.24]{figs/extract_gap_2}}%\n\\subfigure[GapLarge]{%\n\\label{fig:extc1}%\n\\includegraphics[scale=0.24]{figs/extract_gap_3}}%\n\\\\\n\\subfigure[NoiseSmall]{%\n\\label{fig:exta2}%\n\\includegraphics[scale=0.24]{figs/extract_noise_1}}%\n\\subfigure[NoiseMedium]{%\n\\label{fig:extb2}%\n\\includegraphics[scale=0.24]{figs/extract_noise_2}}%\n\\subfigure[NoiseLarge]{%\n\\label{fig:extc2}%\n\\includegraphics[scale=0.24]{figs/extract_noise_3}}%\n\\\\\n\\subfigure[DecaySlow]{%\n\\label{fig:exta3}%\n\\includegraphics[scale=0.24]{figs/extract_decay_1}}%\n\\subfigure[DecayMedium]{%\n\\label{fig:extb3}%\n\\includegraphics[scale=0.24]{figs/extract_decay_2}}%\n\\subfigure[DecayFast]{%\n\\label{fig:extc3}%\n\\includegraphics[scale=0.24]{figs/extract_decay_3}}%\n\\caption{Plots of $\\max\\{\\sin \\theta_j',\\sin\\nu_j'\\} $ for $j=1,\\dots,k$. The test matrices were described in~\\cref{ssec:test}. The target rank $k=15$ and an oversampling parameter of $20$ was chosen for all the experiments. The solid lines correspond to the computed values, the dashed lines correspond to bounds obtained using~\\cref{thm:sintheta}. The parameter $q$ corresponds to the number of subspace iterations.}\n\\label{fig:ext}\n\\end{figure}\n\n\\subsection{Singular Values} We now consider the accuracy of the singular values. We use the same test matrices and the remaining parameters are kept fixed. The computed singular values are plotted against the upper and lower bounds. We make the following general observations:\n{ \n\\begin{itemize}\n\\item For the large singular values, both the upper and lower bounds are qualitatively good for all the examples that we tested.\n\\item As the number of iterations increase, the singular values are computed more accurately and are close to the upper bounds (the exact singular values). However, for indices close to the target rank, the lower bounds are are not tight. The bounds get tighter as the number of iterations $q$ increase. \n\\item The bounds for the singular values quantitatively better than the bounds for the canonical angles. \n\\end{itemize} \n}\n\\begin{figure}[!ht]%\n\\centering\n\\subfigure[GapSmall, $q=0$]{%\n\\label{fig:svsa1}%\n\\includegraphics[scale=0.24]{figs/svs_gap_q_0}}%\n\\subfigure[GapSmall, $q=1$]{%\n\\label{fig:svsb1}%\n\\includegraphics[scale=0.24]{figs/svs_gap_q_1}}%\n\\subfigure[GapSmall, $q=2$]{%\n\\label{fig:svsc1}%\n\\includegraphics[scale=0.24]{figs/svs_gap_q_2}}%\n\\\\\n\\subfigure[NoiseMedium, $q=0$]{%\n\\label{fig:svsa2}%\n\\includegraphics[scale=0.24]{figs/svs_noise_q_0}}%\n\\subfigure[NoiseMedium, $q=1$]{%\n\\label{fig:svsb2}%\n\\includegraphics[scale=0.24]{figs/svs_noise_q_1}}%\n\\subfigure[NoiseMedium, $q=2$]{%\n\\label{fig:svsc2}%\n\\includegraphics[scale=0.24]{figs/svs_noise_q_2}}%\n\\\\\n\\subfigure[DecayMedium, $q=0$]{%\n\\label{fig:svsa3}%\n\\includegraphics[scale=0.24]{figs/svs_decay_q_0}}%\n\\subfigure[DecayMedium, $q=1$]{%\n\\label{fig:svsb3}%\n\\includegraphics[scale=0.24]{figs/svs_decay_q_1}}%\n\\subfigure[DecayMedium, $q=2$]{%\n\\label{fig:svsc3}%\n\\includegraphics[scale=0.24]{figs/svs_decay_q_2}}%\n\\caption{Plots of the singular values. The test matrices were described in \\cref{ssec:test}. The target rank $k=25$ and an oversampling parameter of $\\rho=20$ was chosen for all the experiments. The solid lines black and blue lines correspond to the upper and lower bounds respectively, the dashed red lines correspond to bounds obtained using \\cref{thm:sigma}. The parameter $q$ corresponds to the number of subspace iterations.}\n\\end{figure} \n\n\n\n{\n\nWe now make observations specific to the test examples: \n\\begin{description}\n\\item [1. GapSmall] In these examples, the large singular values are captured accurately. As the number of iterations increase, both the lower bound and the approximate singular values approach the true singular values (upper bound). For GapMedium and GapLarge, the bounds were much more accurate.\n\\item [2. NoiseMedium]  There is a qualitatively different behavior before and after indices $15-16$. The upper and lower bounds are tight before index $15$, but only the upper bound is tight after index $16$. The lower bound significantly under-predicts the singular values.\n\\item [3. DecayMedium] Similar to the previous example, the lower bounds are good before index $15$, and improve with number of iterations after index $15$.\n\\end{description}\n }\n\n\n\n\\section{Acknowledgments}\nThe author would like to thank Ilse C.F.\\ Ipsen and Andreas Stathopoulos for helpful conversations. He would also like to acknowledge Ivy Huang for her help with the figures.\n\n\\iffalse\n", "meta": {"timestamp": "2018-11-13T02:13:48", "yymm": "1804", "arxiv_id": "1804.02614", "language": "en", "url": "https://arxiv.org/abs/1804.02614", "abstract": "This paper is concerned with the analysis of the randomized subspace iteration for the computation of low-rank approximations. We present three different kinds of bounds. First, we derive both bounds for the canonical angles between the exact and the approximate singular subspaces. Second, we derive bounds for the low-rank approximation in any unitarily invariant norm (including the Schatten-p norm). This generalizes the bounds for Spectral and Frobenius norms found in the literature. Third, we present bounds for the accuracy of the singular values. The bounds are structural in that they are applicable to any starting guess, be it random or deterministic, that satisfies some minimal assumptions. Specialized bounds are provided when a Gaussian random matrix is used as the starting guess. Numerical experiments demonstrate the effectiveness of the proposed bounds.", "subjects": "Numerical Analysis (math.NA)", "title": "Randomized subspace iteration: Analysis of canonical angles and unitarily invariant norms", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9833429580381723, "lm_q2_score": 0.8558511524823263, "lm_q1q2_score": 0.8415952039223497}}
{"url": "https://arxiv.org/abs/2102.00996", "title": "Compositions that are palindromic modulo $m$", "abstract": "In recent work, G. E. Andrews and G. Simay prove a surprising relation involving parity palindromic compositions, and ask whether a combinatorial proof can be found. We extend their results to a more general class of compositions that are palindromic modulo $m$, that includes the parity palindromic case when $m=2$. We then provide combinatorial proofs for the cases $m=2$ and $m=3$.", "text": "\\section{Introduction}\n\nLet $\\sigma=(\\sigma_1,\\sigma_2,\\ldots,\\sigma_k)$ be a sequence of positive integers such that $\\sum \\sigma_i = n$. The sequence $\\sigma$ is called a \\textit{composition} of $n$ of length $k$. The numbers $\\sigma_i$ are called the \\textit{parts} of the composition. If $\\sigma_i=\\sigma_{k-i+1}$ for all $i$, then $\\sigma$ is called a \\textit{palindromic} composition. If instead $\\sigma$ satisfies the weaker condition that $\\sigma_i \\equiv \\sigma_{k-i+1}$ modulo $m$ for some $m\\geq 1$ and all $i$, then $\\sigma$ is said to be \\textit{palindromic modulo $m$}. Let $pc(n,m)$ be the number of compositions of $n$ that are palindromic modulo $m$. \n\nAndrews and Simay \\cite{AS20} have shown that $pc(1,2)=1$, and \\[pc(2n,2)=pc(2n+1,2)=2\\cdot 3^{n-1}\\]\nfor $n>1$.\nOur main theorem generalizes their result by giving an ordinary generating function for $pc(n,m)$ for all $m$. Of particular interest is that $pc(1,3)=1$, and \\[pc(n,3)=2\\cdot f(n-1)\\] for $n>1$, where $f(n)$ is the $n$th Fibonacci number (here $f(1)=f(2)=1$). \nWe then give a combinatorial proof of the above identities for $pc(n,2)$ and $pc(n,3)$, and conclude with some general properties and asymptotic analysis of $pc(n,m)$ for larger $m$.\n\n\\begin{thm}\n    We have \\[F_m(q) := \\sum_{n\\geq 1} pc(n,m) q^n = \\frac{q+2q^2-q^{m+1}}{1-2q^2-q^m}. \\] \n\\end{thm}\n\n\\noindent{\\bf Remark.} Though we mean this to be a formal definition, it can be verified that the series converges for all $|q|<1/2$. Indeed, \\[F_m(q) \\leq F_1(q) = \\frac{q}{1-2q}.\\]\n\n\\begin{proof}[Proof of Theorem 1]\n    Notice that to form a composition that is palindromic modulo $m$, we need to form a sequence of pairs of positive integers whose difference is a multiple of $m$. Furthermore, there may or may not be a central part that will always be congruent to itself. If $G_m(q)$ is the ordinary generating function for pairs of positive integers whose difference is a multiple of $m$, then \\[F_m(q) = \\frac{G_m(q)}{1-G_m(q)} \\cdot \\frac{1}{1-q} + \\frac{q}{1-q}.\\] Now for the function $G_m(q)$, either the pair of positive integers is the same, or we must choose to add a multiple of $m$ to either the first or second number. Therefore \\[G_m(q) = \\frac{q^2}{1-q^2}+\\frac{2q^2}{1-q^2}\\cdot \\frac{q^m}{1-q^m}.\\] Simplifying gives the result. \\end{proof}\n    \nWhen $m=1$, every composition will be palindromic modulo $m$. Therefore $pc(n,1)=2^{n-1}$, as is evident from $F_1(q)$. We next consider the case when $m=2$.\n\n\\section{The case $m=2$}\n\nWhen $m=2$, Theorem 1 gives \\[F_2(q) = \\frac{2q^3(1+q)}{1-3q^2},\\] and after expanding we see that $p(1,2)=1$, and $p(2n,2)=p(2n+1,2)=2\\cdot 3^{n-1}$. Our goal in this section will be to give a combinatorial proof of this formula. \n\nFor $n\\geq 1$, we start by considering the $3^{n-1}$ elements of the set $\\{0,1,2 \\}^{n-1}$. Our goal will be to embed two disjoint copies of this set into the set of compositions of $2n$ that are palindromic modulo 2, and show that every composition of $2n$ that is palindromic modulo 2 has a preimage. We then give a bijection between compositions of $2n$ that are palindromic modulo 2 and compositions of $2n+1$ that are palindromic modulo 2.\n\n\\begin{proof}[Combinatorial proof that $pc(2n,2)=2\\cdot3^{n-1}$] For $n\\geq 1$, let \\[a=(a_1,a_2,\\ldots,a_{n-1})\\in \\{0,1,2 \\}^{n-1}.\\] As an example, we take $a=(0,1,2,2,1)\\in \\{0,1,2 \\}^5$. We will first use the sequence of $a_i$s to construct a sequence of triples $(b_j,c_j,d_j)$, where each triple will become a pair of  parts $\\{ \\sigma_j,\\sigma_{k-j+1}\\}$ or the central part $\\sigma_{(k+1)/2}$ in the resulting composition having length $k$.\n\nInitialize the process with the triple $(b_1,c_1,d_1)=(1,0,0)$. If $n=1$, then we are done. Now if $a_1=0$, we create a new triple $(b_2,c_2,d_2)=(1,0,0)$. If $a_1=1$, then we increase $b_1$ by one. If $a_1=2$, then we increase $c_1$ by one. In our example of $(0,1,2,2,1)$, we see that $a_1=0$, so after the first step we have the following list of triples\n\\[(1,0,0),  \\ \\ (1,0,0).\\]\n\nNow for $i>1$, we define the instructions given by each $a_i$ recursively. We assume that at this step we have a total of $j$ triples, for some $j\\geq 1$. \n\\begin{enumerate}\n    \\item If $a_i=0$, then we create the new triple $(b_{j+1},c_{j+1},d_{j+1})=(1,0,0)$.\n    \\item If $a_i=2$, then increase $c_j$ by one.\n    \\item If $a_i=1$ and $a_{i-1}\\neq 2$, then increase $b_j$ by one. \n    \\item If $a_i=1$ and $a_{i-1}=2$, then set $d_j=1$ and create the new triple $(b_{j+1},c_{j+1},d_{j+1})=(1,0,0)$.\n\\end{enumerate}\n   For our example $(0,1,2,2,1)$, we form the list of triples\n\\[(1,0,0), \\ \\ (2,2,1), \\ \\  (1,0,0). \\]\n\nAfter performing this algorithm on each of the $3^{n-1}$ sequences in $\\{ 0,1,2 \\}^{n-1}$ gives a set of $3^{n-1}$ sequences of triples. For each sequence $a\\in\\{0,1,2\\}^{n-1}$, we denote by $k_a$ the number triples in the list we have formed. By construction, all of these sequences of $k_a$ triples are distinct.\n\nWe create a second set containing $3^{n-1}$ sequences of triples by setting $d_{k_a} = 1$ for all $a$. For our example $(0,1,2,2,1)$, we get a second sequence of triples\n\\[(1,0,0), \\ \\ (2,2,1), \\ \\  (1,0,1). \\]\nWe now have $2\\cdot 3^{n-1}$ sequences of triples, all of which are distinct.\n\nNext, for each sequence $a$, taking one of its two associated sequence of triples $(b_1,c_1,d_1)$, \\ldots, $(b_{k_a},c_{k_a},d_{k_a})$, we show how to form a unique composition $\\sigma$ of $2n$ that is palindromic modulo 2. Start with the triple $(b_{k_a}, c_{k_a}, d_{k_a})$. \n\\begin{enumerate}\n    \\item If $d_{k_a}=0$ and $c_{k_a}=0$, then set $\\sigma_{k_a}=2b_{k_a}$.\n    \\item If $d_{k_a}=0$ and $c_{k_a}>0$, then set $\\sigma_{k_a}=b_{k_a}+2c_{k_a}$ and $\\sigma_{k_a+1}=b_{k_a}$.\n    \\item If $d_{k_a}=1$, then set $\\sigma_{k_a}=b_{k_a}$ and $\\sigma_{k_a+1}=b_{k_a}+2c_{k_a}$.\n\\end{enumerate}\n  These cases will determine whether the length of the composition is even or odd. If $d_{k_a}=0$ and $c_{k_a}=0$, then the length of $\\sigma$ will be $2k_a-1$. Otherwise, the length of $\\sigma$ will be $2k_a$. For convenience we refer to the length of $\\sigma$ as $k$ in either case. For our example $(0,1,2,2,1)$, we have initialized the two compositions \\[(0,0,2,0,0) \\ \\text{and} \\ (0,0,1,1,0,0).\\]\n\nNow if $k_a=1$ we are done. Assuming $k_a>1$, for each $1\\leq j < k_a$ we consider the triple $(b_j,c_j,d_j)$. If $d_j=0$, then set $\\sigma_j=b_j+2c_j$ and $\\sigma_{k-j+1} = b_j$. If $d_j=1$, then set $\\sigma_j=b_j$ and $\\sigma_{k-j+1}=b_j+2c_j$. Also notice if $d_j=1$, then by construction $c_j>0$, so that $b_j\\neq b_j+2c_j$. This ensures that each sequence of triples will be associated to a unique composition. Also notice that the sum of the parts in $\\sigma$ equals $2\\sum b_j + 2\\sum c_j=2n$, and that $|\\sigma_j-\\sigma_{k-j+1}|$ is even for all $j$. Thus $\\sigma$ is a composition of $2n$ that is palindromic modulo 2. Showing that every composition of $2n$ that is palindromic modulo 2 can be formed from one of the described sequences of triples is done by reversing the described algorithm. Returning to our example sequence $(0,1,2,2,1)$, we have form the two compositions\n\\[(1,2,2,6,1) \\ \\text{and} \\ (1,2,1,1,6,1),\\]\neach of which is a composition of 12 that is palindromic modulo 2.\\end{proof}\n\n\\begin{proof}[Combinatorial proof that $pc(2n,2)=pc(2n+1,2)$]\n    We split the proof into two cases. Starting with a composition $\\sigma$ of $2n$ that is palindromic modulo 2, suppose the length of $\\sigma$ is $2k+1$. Then adding one to $\\sigma_{k+1}$ gives a composition of $2n+1$ that is still palindromic modulo 2. Now if the length of $\\sigma$ is $2k$, form a composition $\\sigma^{\\prime}$ of $2n+1$ of length $2k+1$ by setting \n    \\[\\sigma^{\\prime}_j=\\begin{cases}\n        \\sigma_j & 1\\leq j\\leq k \\\\\n        1 & j=k+1 \\\\\n        \\sigma_{j-1} & k+2\\leq j \\leq 2k+1.\n    \\end{cases}\n    \\]\nWe note that $\\sigma^{\\prime}$ is still palindromic modulo 2. It is straightforward to verify that this map is a bijection. For our examples of $(1,2,2,6,1)$ and $(1,2,1,1,6,1)$ that are compositions of 12, we form the two compositions of 13 \n\\[(1,2,1,1,1,6,1) \\ \\ \\text{ and } \\ \\ (1,2,3,6,1),\\] \neach of which is palindromic modulo 2. \\end{proof}\n\n\\section{The case $m=3$}\n\nWhen $m=3$, Theorem 1 gives \n\\[F_3(q) = \\frac{q+2q^2-q^4}{1-2q^2-q^3} = q + \\frac{2q^2}{1-q-q^2},\\]\nand expanding this function shows $pc(1,3)=1$ and $pc(n,3)=2\\cdot f(n-1)$ for $n>1$.  Our goal of this section is to give a combinatorial proof of this formula.\n\nIt is well known \\cite{AH75} that $f(n+1)$ is equal to the number of compositions of $n$ with parts that are equal to one or two. For $n>1$, our goal will be to embed two disjoint copies of the compositions of $n-2$ with parts equal to one or two into the compositions of $n$ that are palindromic modulo 3. Then we will show each composition of $n$ that is palindromic modulo 3 has a preimage. \n\n\\begin{proof}[Combinatorial proof that $pc(n,3)=2\\cdot f(n-1)$.] For $n>1$, let $\\sigma =( \\sigma_1,\\sigma_2,\\ldots,\\sigma_k)$ be a composition of $n-2$, where each $\\sigma_i\\in \\{1,2\\}$. Form two distinct compositions of $n$ by setting \n\\begin{align*}\n    \\sigma^{\\prime}&=(1,1,\\sigma_1,\\sigma_2,\\ldots,\\sigma_k) \\text{, and} \\\\\n    \\sigma^{\\prime\\prime}&=(2,\\sigma_1,\\sigma_2,\\ldots,\\sigma_k).\n\\end{align*}\nDoing this for each composition gives us a set of $2\\cdot f(n-1)$ compositions of $n$ with parts equal to one or two, the set of which we will denote by $A_n$. For each of these compositions, we will form a unique composition of $n$ that is palindromic modulo 3.\n\nLet $+$ denote sequence concatenation, as in $(1,2)+(3,4)=(1,2,3,4)$. Now for any composition $a\\in A_n$, we can decompose $a$ in the following way:\n\\[a=a_1+a_2+\\ldots+a_s,\\]\nwhere the substrings $\\{a_i\\}_{i=1}^s$ are determined by the following rules. We will write $k_i$ to be the length of each substring $a_i$.\n\n\\begin{enumerate}\n    \\item If $a$ contains no twos, then $a_1=a$.\n    \\item Suppose $a$ begins with a two, and this is the only two in $a$. Then $a_1=(2)$.\n    \\item Suppose $a$ begins with the substring $(2,2)$ or $(2,1,1)$. Then $a_1=(2)$.\n    \\item Suppose $a$ begins with the substring $(1,1)$ or $(2,1,2)$. Then $a_1$ terminates with the first two such that\n    \\begin{enumerate}\n        \\item it is the final two in $a$;\n        \\item it is immediately followed by the substring $(2)$ or $(1,1)$.\n    \\end{enumerate}\n   \\end{enumerate}\nFor $i>1$, the substring $a_i$ is determined recursively using these 4 rules, after first deleting the substrings $a_1$, $a_2$, \\ldots, $a_{i-1}$ from $a$. The resulting string cannot begin with the substring $(1,2)$, so it still makes sense to apply these rules. To aid the reader, Table \\ref{decomp} shows the decomposition of the compositions in the set $A_8$. As an example, take the composition $a=(1,1,1,2,2,1,1,2,1,2,1,1)$ in $A_{16}$. By applying rule (4) we have $a_1=(1,1,1,2)$, and we now consider the string $(2,1,1,2,1,2,1,1)$. By applying rule (3) we have $a_2=(2)$, and we now consider the string $(1,1,2,1,2,1,1)$. By applying rule (4) we have $a_3=(1,1,2,1,2)$, and we now consider the string $(1,1)$. By applying rule (1) we have $a_4=(1,1)$, and are now done. Thus\n\\[(1,1,1,2,2,1,1,2,1,2,1,1)=(1,1,1,2)+(2)+(1,1,2,1,2)+(1,1).\\]\n\n\nLet $B_n$ be the set of compositions of $n$ that are palindromic modulo 3. For each composition $a=a_1+\\ldots+a_s\\in A_n$, we will show how to construct a unique $b\\in B_n$. If $a_s$ contains a two, the length of $b$ will be $2s$; if $a_s$ does not contain a two, the length of $b$ will be $2s-1$. Either way, we will denote $k_b$ to be the length of $b$ and write $b=(b_1,b_2,\\ldots,b_{k_b})$ (recall we have also set $k_i$ to be the length of $a_i$).\n\nNow assume $a_i$ has $o_i$ ones, and $t_i$ twos. If $t_i=0$ (which can only be the case for $a_s$), then set $b_s=k_s$. For what follows we will assume $t_i>0$. Note that $k_i=o_i+t_i$, and \\[n=\\sum_{i=1}^s (o_i+2\\cdot t_i).\\]\nWe will now form the triple $(c_i,d_i,e_i)$ using the following rules.\n\nLet $o^{\\prime}_i$ be the number of ones in $a_i$ preceding the first two. Then if\n\\begin{enumerate}\n    \\item $o^{\\prime}_i$ is even (possibly zero), set \n    \\[c_i=\\frac{o^{\\prime}_i}{2}+1, \\ \\ d_i=3\\cdot(t_i-1), \\ \\ \\text{and} \\ \\ e_i=0. \\]\n    \\item $o^{\\prime}_i$ is odd, set \n    \\[c_i=\\frac{o^{\\prime}_i-1}{2}, \\ \\ d_i=0, \\ \\ \\text{and} \\ \\ e_i=3\\cdot t_i. \\]\n\\end{enumerate}\n\nBy construction this sequence of triples is unique to the composition $a\\in A_n$ we began with. We show how our example composition with decomposition $a_1=(1,1,1,2)$, $a_2=(2)$, $a_3=(1,1,2,1,2)$, and $a_4=(1,1)$ maps to three triples (the final substring $a_4$ contains no twos, and therefore we set $b_4=k_4=2$).\n\\begin{center}\n    \\begin{tabular}{c||c|c|c|c|c|c}\n         $a_i$ & $o_i$ & $o^{\\prime}_i$ & $t_i$ & $c_i$ & $d_i$ & $e_i$  \\\\\n         \\hline\n         $(1,1,1,2)$    &3&3&1&1&0&3\\\\\n         $(2)$          &0&0&1&1&0&0\\\\\n         $(1,1,2,1,2)$  &3&2&2&2&3&0\n    \\end{tabular}\n\\end{center} \nTherefore, we have the triples\n\\[(1,0,3), \\ \\ (1,0,0), \\ \\ \\text{ and } \\ \\ (2,3,0).\\]\n\n\nWe now form $b$ by setting $b_i=c_i+d_i$ and $b_{k_b-i+1}=c_i+e_i$, adjusting slightly for the case when $a_s$ has no twos (recall that we just set $b_s=k_s$ in this case). The composition is uniquely determined from the triples, and by construction $d_i$ and $e_i$ are both multiples of 3. Furthermore, \n\\[b_i+b_{k_b-i+1} = 2c_i + d_i+e_i = o^{\\prime}_i +3t_i -1=o_i+2t_i. \\]\nTo see why the last equality holds, consider first the case when $t_i=1$. Then $o^{\\prime}_i=o_i$, and $3t_i-1=2t_i$. Now if $t_i>1$, we can pair each two in $a_i$ with a one immediately preceding it. However, we have used a one counted by $o^{\\prime}_1$ in this pairing, so we must subtract one. Therefore, we have embedded $A_n$ into $B_n$. For our example\n\\[a=(1,1,1,2,2,1,1,2,1,2,1,1)=(1,1,1,2)+(2)+(1,1,2,1,2)+(1,1),\\]\nwe have the triples $(1,0,3)$, $(1,0,0)$, and $(2,3,0)$, and we obtain the composition\n\\[b=(1,1,5,2,2,1,4),\\]\nwhich is a composition of 16 that is palindromic modulo 3 ($b\\in B_{16}$). It is straightforward to construct a composition $a\\in A_n$ from a composition $b\\in B_n$ by reversing this construction, which proves the result. \\end{proof}\n\n\n\n\n\n\\section{The case $m>3$}\n\nIf $m>3$, the formula for $pc(n,m)$ can be deduced from Theorem 1; we only give a detailed study for $m=2$ and $m=3$ based on the elegance of the formulae. For $m=4$, Theorem 1 gives \n\\[F_4(q)=\\frac{q+2q^2-q^5}{1-2q^2-q^4},\\]\nand after expanding we see $pc(1,4)=1$, and for $n\\geq 1$\n\\[pc(2n,4)=pc(2n+1,4)=\\frac{(1+\\sqrt{2})^n-(1-\\sqrt{2})^n}{\\sqrt{2}}.\\]\nThe familiar reader will recognize these numbers as twice the Pell numbers \\cite{pell}. While the correct context for a combinatorial proof is not immediately clear, we pose the following question.\n\n\\begin{question}\n    Is there a combinatorial proof of the formula for $pc(n,4)$?\n\\end{question}\n\nWe also include some properties of $pc(n,m)$ that we observed for general $m$.\n\n\\begin{prop}\n    If $m$ is even, then for $n\\geq 1$ we have $pc(2n,m)=pc(2n+1,m)$.\n\\end{prop}\n\n\\begin{proof}\n    The proof of this is identical in spirit to the combinatorial proof that $pc(2n,2)=pc(2n+1,2)$ in Section 2. \\end{proof}\n\n\\begin{prop}\n    If $n>1$, then $pc(n,m)$ is even for all $m$.\n\\end{prop}\n\n\\begin{proof}\n    A quick, one line proof of this fact comes from the identity\n    \\[F_m(q)-q = \\frac{2q^2 (1+q)}{1-2q^2-q^m}.\\]\n    This can also be seen combinatorially by pairing up compositions of $n$ that are palindromic modulo $2$. Let $a$ be a composition of $n>1$ of length $s$ that is palindromic modulo $m$. If $a_i\\neq a_{s-i+1}$ for some $i$, then we can pair this composition with the composition formed by switching $a_i$ and $a_{s-i+1}$. \n    \n    Now assume $a_i=a_{s-i+1}$ for all $i$ (i.e. $a$ is a palindrome). If $s$ is odd, than $a_{(s+1)/2}$ is even, and we can pair this composition with the composition of $n$ of length $s+1$ formed by removing $a_{(s+1)/2}$ and appending $a_{(s+1)/2}/2$ to the beginning and the end. \\end{proof}\n\n\\begin{prop}\n    If $2n\\geq m$, then $p(2n,m)=p(2n+1,m)=2^n$.\n\\end{prop}\n\n\\begin{proof}\n    The condition that $2n\\geq m$ requires all compositions of $2n$ and $2n+1$ that are palindromic modulo $m$ to be palindromes. Let $pc(n,\\infty)$ denote the number of palindromic compositions of $n$. Then it is well known \\cite{wk} that\n    \\[pc(n,\\infty)=2^{\\lfloor n/2\\rfloor}. \\qedhere\\]\n\\end{proof}\n\n\\begin{prop}\n    Let $\\alpha_m$ be the unique positive root of $1-2q^2-q^m=0$, and set \n    \\[c_m=\\lim_{q\\rightarrow \\alpha_m} (1-\\alpha_m^{-1}q)\\cdot F_m(q) \\ \\ \\text{ and } \\ \\ d_m=\\lim_{q\\rightarrow \\alpha_m} (1+\\alpha_m^{-1}q)\\cdot F_m(q).\\]\n   If $m$ is even, then\n    \\[\\limsup_{n\\rightarrow \\infty} \\alpha_m^n\\cdot pc(n,m) = c_m+d_m \\ \\ \\text{ and } \\ \\ \\liminf_{n\\rightarrow\\infty} \\alpha_m^n\\cdot pc(n,m)= c_m-d_m. \\]\n    If $m$ is odd, $d_m=0$, and thus $pc(n,m)\\sim c_m \\alpha_m^{-n}$ as $n\\rightarrow \\infty$.\n\\end{prop}\n\n\\begin{proof}\n    This is a routine analysis of the rational function $F_m(q)$. For convenience, we set $\\phi_m(q)=1-2q^2-q^m$. The asymptotics of $pc(n,m)$ are determined by the poles of $F_m(q)$, which are precisely the zeros of $\\phi_m(q)$. We will assume $m>4$, as the result can be verified directly for $m\\leq 4$. \n    \n    First note that $\\phi_m(q)$ has exactly one positive zero, $\\alpha_m$, since $\\phi_m^{\\prime}(q)<0$ for all $q>0$, $\\phi_m(0)>0$, and $\\phi_m(1)<0$. When $m$ is even, $\\phi_m(q)$ is an even function and thus $-\\alpha_m$ is also a zero. When $m$ is odd, $\\phi_m(q)$ has a unique zero in the interval $(-1,-\\alpha_m)$, since $\\phi^{\\prime}_m(q)>0$ on $(-1,0)$, $\\phi_m(-1)<0$, and $\\phi_m(\\alpha_m)>0$. Our next goal is to show that these are the only two zeros inside the disc $|q|<0.9$.\n    \n    Note that $q^2(2+q^{m-2})$ has one zero (with multiplicity equal to 2) inside the disc $|q|<0.9$. Therefore, if we can show\n    \\[|q^2(2+q^{m-2}|>1\\]\n    for all $q$ with $|q|=0.9$, we can apply Rouche's theorem to conclude that $\\phi_m(q)$ has exactly two zeros inside the disc $|q|<0.9$. For $m>4$ we have $|q^{m-2}|\\leq |q^3|$, so that $|2+q^{m-2}|\\geq 2-|q^3|=1.279$ and\n    \\[|q^2(2+q^{m-2})| > 0.81\\cdot 1.729 = 1.03599>1.\\]\n    \n    We can now perform a partial fraction decomposition $F_m(q)$,\n    \\[F_m(q) = \\frac{c_m}{1-\\alpha_m^{-1}q} + \\frac{d_m}{1+\\alpha_m^{-1}q} +G(q),\\]\n    where $G(q)$ is a rational function with no poles in the disc $|q|<0.9$. Therefore,\n    \\[pc(n,m) = (c_m + (-1)^n d_m)\\alpha_m^{-n} + o(\\alpha_m^{-n}) \\]\n    as $n\\rightarrow \\infty$, noting that $d_m=0$ if $m$ is odd. The result follows. \\end{proof}\n    \n    The following table shows the values of $\\alpha_m$, $c_m$, and $d_m$ for various $m$.\n    \\begin{center}\n        \\begin{tabular}{c||c|c|c}\n             $m$&$\\alpha_m^{-1}$&$c_m$&$d_m$  \\\\\n             \\hline\n             1& $2$&$\\frac{1}{2}=0.5$& 0 \\\\\n             2& $\\sqrt{3}\\approx 1.73$&$\\frac{3+\\sqrt{3}}{9}\\approx0.53$& $\\frac{3-\\sqrt{3}}{9}\\approx 0.14$ \\\\\n             3& $\\frac{\\sqrt{5}+1}{2}\\approx 1.61$&$\\frac{5-\\sqrt{5}}{5}=0.55$& 0 \\\\\n             4& $\\approx 1.55$&$\\approx0.58$& $\\approx 0.13$ \\\\\n             \\vdots&\\vdots & \\vdots & \\vdots \\\\\n             $\\infty$ &$\\sqrt{2}\\approx 1.41$&$\\frac{2+\\sqrt{2}}{4}\\approx 0.85$& $\\frac{2-\\sqrt{2}}{4}\\approx 0.14$\n        \\end{tabular}\n    \\end{center}\n\n\n\\section*{Acknowledgements}\n\nThe author was partially supported by the Research and Training Group grant DMS-1344994 funded by the National Science Foundation.\n\n", "meta": {"timestamp": "2021-09-29T02:08:40", "yymm": "2102", "arxiv_id": "2102.00996", "language": "en", "url": "https://arxiv.org/abs/2102.00996", "abstract": "In recent work, G. E. Andrews and G. Simay prove a surprising relation involving parity palindromic compositions, and ask whether a combinatorial proof can be found. We extend their results to a more general class of compositions that are palindromic modulo $m$, that includes the parity palindromic case when $m=2$. We then provide combinatorial proofs for the cases $m=2$ and $m=3$.", "subjects": "Combinatorics (math.CO)", "title": "Compositions that are palindromic modulo $m$", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9924227585080309, "lm_q2_score": 0.8479677602988602, "lm_q1q2_score": 0.8415425038016715}}
{"url": "https://arxiv.org/abs/1505.04151", "title": "Minkowski Symmetrizations of Star Shaped Sets", "abstract": "We provide sharp upper bounds for the number of symmetrizations required to transform a star shaped set in ${\\mathbb R}^n$ arbitrarily close (in the Hausdorff metric) to the Euclidean ball.", "text": "\\section{Introduction and results}\\label{Sec_Intro}\nA non empty compact set $K\\subset {\\cal R}^n$ is called {\\em star shaped}\nif $x\\in K$ implies $[0,x]\\subseteq K$. We denote the family of star\nshaped sets in ${\\cal R}^n$ by $\\mathscr{S}^n$. Recall that given a set $K$ and a\ndirection $u \\in S^{n-1}$, it's Minkowski symmetral is defined to be\n\\[\nM_u(K) = \\frac{K + R_u K}{2},\n\\] \nwhere $R_u$ is the reflection with respect to the hyperplane\n$u^\\perp$.\n\nThe Minkowski symmetrization $M_u$ results in a set that is symmetric\nwith respect to the hyperplane $u^\\perp$, thus it is natural to\nexpect that successive applications of this procedure in different\ndirections yield a sequence of sets that convergences in some sense\nto the Euclidean ball.\nThis is indeed known in the case where $K$ is convex.\nMoreover, there are estimates regarding the convergence rate.\nBourgain, Lindenstrauss and Milman \\cite{BLM} obtained the first\nquantitative estimate for the convergence rate of Minkowski\nsymmetrizations. They found a function $n_0: (0,1)\\to \\mathbb N$ satisfying:\n\\begin{thm} [Bourgain, Lindenstrauss, Milman] \\label{thm-Intro-BLM}\nLet $\\varepsilon \\in(0,1)$ and let $n\\in \\mathbb N$ such that\n$n_0(\\varepsilon)\n\\le n$. If $K\\in {\\cal K}^n$ is a convex body with mean width $M^*(K)=M_0$,\nthen there exist $cn\\left(C(\\varepsilon) + \\log n\\right)$ Minkowski\nsymmetrizations transforming $K$ into a (convex) body $\\tilde{K}$,\nsuch that\n\\[\n(1-\\varepsilon) M_0 D_n \\subset\n\\tilde{K} \\subset\n(1+\\varepsilon) M_0 D_n.\n\\]\nHere $c$ is some positive constant, and the functions\n$C(\\varepsilon)$, $n_0(\\varepsilon)$ are of the order\n$\\left(\\frac{1}{\\varepsilon}\\right)^\\frac{c}{\\varepsilon^2}$.\n\\end{thm}\n\nIn \\cite{K_Rate} Klartag improved Theorem \\ref{thm-Intro-BLM}, and\nalso removed the restriction $n_0\\le n$, thus providing the first\ntruely isometric result\nin all dimensions:\n\\begin{thm}[Klartag] \\label{thm-Klartag}\nLet $n\\ge 2$ and $\\varepsilon\\in(0,1/2)$. If $K\\in {\\cal K}^n$ is a convex\nset with mean width $M^*(K)=M_0$, then there exist $cn |\\log\n(\\varepsilon)|$ Minkowski symmetrizations transforming $K$ into\n$\\tilde{K}$, such that\n\\[\n(1-\\varepsilon)M_0 D_n\n\\subseteq \\tilde{K} \\subseteq\n(1+\\varepsilon)M_0 D_n,\n\\] where $c$ is some universal constant.\n\\end{thm}\n\nIn this note we extend Klartag's theorem to $\\mathscr{S}^n$. Namely, we show the following:\n\n\\begin{thm}\\label{Thm_Mink-Isometric}\nLet $n\\ge 2$ and $\\varepsilon\\in(0,1/2)$. If $K\\in \\mathscr{S}^n$ is a star\nshaped set with mean width $M^*(K)=M_0$, then there exist $Cn\n|\\log(\\varepsilon)|$\nMinkowski symmetrizations transforming $K$ into $\\tilde{K}$, such that\n\\[\n(1-\\varepsilon)M_0 D_n\n\\subseteq \\tilde{K} \\subseteq\n(1+\\varepsilon)M_0 D_n,\n\\] where $C$ is some universal constant.\n\\end{thm}\n\nThe proof of Theorem \\ref{Thm_Mink-Isometric} consists of three steps.\nFirst, we make sure that the body at hand contains a small ball around\nthe origin (Lemma \\ref{Lem_Give-Small-Ball\n). Next, we consider Minkowski symmetrizations {\\em of the convex\n\thull} of our star shaped body. As mentioned above, for convex bodies\nKlartag showed how many steps are required in order to bring a body\nisometrically close to the Euclidean ball. We apply these\nsymmetrizations, and get a body whose convex hull lies between two\nballs of very similar radii. This can only happen if the body\ncontains some ``$\\varepsilon$-net'' of the inner ball's sphere (Lemma\n\\ref{Lem_Conv=Ball-to-eps-net}). In the third and final step, we use\nthis fact to increase the radius of the small ball, which was\nobtained in the first step.\n\n\n\\noindent {\\bf Notations:}\nThe {\\em support function} of a (not necessarily convex) body $K$ is\ndefined by $h_K(u) = \\sup \\{ \\iprod{x}{u} \\ | \\  x \\in  K \\}$. The\nwidth, or {\\em mean width}, of a star shaped set $K$ is defined to be\n$M^*(K)= \\int_{S^{n-1}} h_K d\\sigma$, where $\\sigma$ is the\nnormalized Haar measure on the sphere.\n\n\n\\section{Proof of the Theorem\n}\\label{Sec_Mink}\n\nOur first step is to generate (using Minkowski symmetrizations) a\nsmall ball inside a (non trivial) star shaped set.\n\n\\begin{lem}\\label{Lem_Give-Small-Ball}\nLet $n\\ge 2$, $K_0\\in \\mathscr{S}^n$, and let $M_0=M^*(K_0)>0$. Then there exist\n$c_1n$ Minkowski symmetrizations transforming $K_0$ into $K_1$, such that\n\\[\n\\frac{c_2}{\\sqrt{n}} D_n\n\\subseteq \\frac{K_1}{M_0}\n,\\]\nwhere $c_1,c_2$ are some universal constants (in fact $c_1=c$ of\nTheorem \\ref{thm-Klartag}).\n\\end{lem}\n\\begin{proof}\nLet $R\\ge M_0>0$ be the minimal radius of a centered ball enclosing\n$K_0$. Then there exists some $u\\in S^{n-1}$ such that $I_0=[0,Ru]\n\\subseteq K_0$. Let $\\varepsilon = 1/e$. By Theorem \\ref{thm-Klartag}\napplied to $I_0$, there exist $N_1 = cn$\nMinkowski symmetrizations\n$M_{u_1}\\dots M_{u_{N_1}}$ which transform the interval $I_0$ into a\nconvex body $I_1$ satisfying\n\\[\n(1-1/e) M^*(I_0) D_n \\subseteq I_1.\n\\]\nThen\n\\[\n(1-1/e) M^*([0,u]) D_n\n\\subseteq\n(1-1/e) \\left(\\frac{RM^*([0,u])}{M_0}\\right) D_n\n\\subseteq\n\\frac{I_1}{M_0}\n\\subseteq\n\\frac{K_1}{M_0},\n\\]\nwhere the body $K_1$ is defined by:\n\\begin{equation}\\label{Eq_Mink-1st-Step}\nK_1:=M_{u_{N_1}}\\dots M_{u_1} K_0.\n\\end{equation}\nSince $M^*([0,u])=\n\\frac{1}{2}\\int_{S^{n-1}}|x_1|d\\sigma(x)\\approx\n\\frac{1}{\\sqrt{2\\pi n}}$, the proof is complete.\n\\end{proof}\n\nThe inner radius $c_2/\\sqrt{n}$ does not decrease under additional\nsymmetrizations (it may increase). Next consider Minkowski\nsymmetrizations which bring {\\em the convex hull} of $K_1$\nisometrically close to the ball. By Theorem \\ref{thm-Klartag} there\nexist $N_2= c n |\\log(\\varepsilon)|$\ndirections $u_1,\\dots,u_{N_2}$\nsuch that\n\\[\n(1 - \\varepsilon) D_n\n\\subseteq\nM_{u_{N_2}}\\dots M_{u_1} {\\rm conv} \\left( \\frac{K_1}{M_0} \\right)\n\\subseteq\n(1 + \\varepsilon) D_n.\n\\]\nWe define the body $K_2$ to be:\n\\begin{equation}\\label{Eq_Mink-2nd-Step}\nK_2:=M_{u_{N_2}}\\dots M_{u_1} K_1.\n\\end{equation}\n\nNote that Minkowski symmetrizations commute with the convex hull\noperation, i.e. $M_u{\\rm conv} K = {\\rm conv} M_u K$, simply because in general\n${\\rm conv} (A+B) = {\\rm conv} A + {\\rm conv} B$. Thus $\\frac{K_2}{M_0}$ is a star\nshaped body whose convex hull is isometrically close to the ball. We\nuse the following standard lemma to show that such a body must\ncontain some $\\delta$-net of the sphere. More precisely:\n\\begin{lem}\\label{Lem_Conv=Ball-to-eps-net}\nLet $n\\ge 2$, $\\varepsilon\\in(0,1)$, and let $K\\in \\mathscr{S}^n$ be such that\n\\[(1 - \\varepsilon) D_n\n  \\subseteq  {\\rm conv} K  \\subseteq\n  (1 + \\varepsilon) D_n.\n\\]\nThen $K$ contains a $2\\sqrt{\\varepsilon}$-net of the sphere\n$(1-\\varepsilon) S^{n-1}$, that is\n\\begin{equation}\\label{Eq_Delta-net}\n(1-\\varepsilon) S^{n-1}\n\\subseteq\nK + 2\\sqrt{\\varepsilon}D_n.\n\\end{equation}\n\\end{lem}\n\n\\begin{proof}\nLet $x\\in (1-\\varepsilon) S^{n-1}$. We claim that the intersection\n$ (x+2\\sqrt{\\varepsilon}D_n)\\cap K $ is not empty. Denoting the\nhyperplane supporting $(1-\\varepsilon) D_n$ at the point $x$ by $H$\nand the halfspace with boundary $H$ by $H^+= \\{y\\,|\\,\n\\iprod{x}{x} \\le \\iprod{x}{y} \\}$, we will show that $H^+\\cap K \\cap\n(x+2\\sqrt{\\varepsilon}D_n)$ is not empty. Indeed,\n\\[\nH^+ \\cap K\n\\subseteq\nH^+ \\cap (1+\\varepsilon) S^{n-1}\n\\subseteq\n(x + 2\\sqrt{\\varepsilon}D_n),\n\\]\nsee Figure \\ref{Fig_Instead-of-Proving}. The set $H^+ \\cap K$ is not\nempty (since $x\\in (H^+ \\cap{\\rm conv} K)$), and thus the proof is complete.\n\\begin{figure}[h]\n\t\\centering\n\t\\includegraphics[width=0.35\\textwidth]{conv_is_ball}\n\t\\caption{The small ball around $x$ contains all points of $K$ in $H^+$.}\n\t\\label{Fig_Instead-of-Proving}\n\\end{figure}\n\\end{proof}\n\nNote that in fact, since $K + 2\\sqrt{\\varepsilon}D_n$ is star shaped, we have:\n\\begin{equation}\n(1-\\varepsilon) D_n\n\\subseteq\nK + 2\\sqrt{\\varepsilon}D_n.\n\\end{equation}\nThe outer radius $1+\\varepsilon$ does not increase under additional\nsymmetrizations (it may decrease). As for the inner radius, we begin\nwith the small ball obtained in Lemma \\ref{Lem_Give-Small-Ball}, and\nuse \\eqref{Eq_Delta-net} to increase it geometrically. More precisely:\n\n\\begin{lem} \\label{lem-StarShapedContainsBall}\nLet $n\\ge 2$, $K\\in \\mathscr{S}^n$, $r\\in(0,1)$, and $\\varepsilon\\in\n(0,\\varepsilon_0)$, where $\\varepsilon_0=1/25$. Assume that:\n\\begin{itemize}\n\\item $r D_n \\subseteq K$.\n\\item $(1-\\varepsilon) D_n\n\\subseteq\nK + 2\\sqrt{\\varepsilon}D_n$.\n\\end{itemize}\nThen there exist $N=\\alpha + \\beta|\\log \\varepsilon| + \\gamma|\\log r|$ Minkowski symmetrizations transforming $K$\ninto $\\tilde{K}$ satisfying\n\\[\n(1-4\\sqrt{\\varepsilon}) D_n \\subseteq \\tilde{K},\n\\]\nwhere $\\alpha, \\beta, \\gamma$ are positive constants.\n\\end{lem}\n\\begin{proof}\nIn each of the two cases $r < 2\\sqrt{\\varepsilon}$ and\n$2\\sqrt{\\varepsilon} \\le r$, we argue a bit differently, so we handle\nthem separately. We begin with the case of smaller initial inner radius.\n\n\\noindent{\\bf Case a. Increasing $r$ geometrically to reach $2\\sqrt{\\varepsilon}$:}\n\nIf $r < 2\\sqrt{\\varepsilon}$ we may take the second assumption and\nwrite for any $u\\in S^{n-1}$:\n\\[\nr\\frac{1-\\varepsilon}{2\\sqrt{\\varepsilon}} D_n\n\\subseteq\n\\frac{r}{2\\sqrt{\\varepsilon}}K + r D_n\n\\subset\nK + r D_n\n\\subseteq\nK + R_u (K),\n\\]\nSo that $r\\frac{1-\\varepsilon}{4\\sqrt{\\varepsilon}} D_n\n\\subseteq M_u (K)$. Defining the (decreasing) function $q:(0,1)\\to{\\cal R}^+$\nby $q(\\varepsilon)= \\frac{1-\\varepsilon} {4\\sqrt{\\varepsilon}}$, we\nhave $q(\\varepsilon)\\ge q(\\varepsilon_0)= 6/5$, so the inner radius\nmultiplies by at least $6/5$. If $2\\sqrt{\\varepsilon} \\le r(6/5)^m$,\nthen after $m$ symmetrizations the inner radius reaches\n$2\\sqrt{\\varepsilon}$.\nThus after $N_a = 4 + 3 \\left|\\log\\left( \\frac{\\varepsilon}{r^2} \\right)\\right|$\nsymmetrizations, we have reduced to the second case, where\n$2\\sqrt{\\varepsilon} \\le r$.\n\n\\noindent{\\bf Case b. Increasing $r$ geometrically  towards $1$:}\n\nAgain, for any $u\\in S^{n-1}$ we have\n\\[\n\\frac{(1-[2\\sqrt{\\varepsilon} + \\varepsilon]) + r}{2} D_n\n=\n\\frac{(1 - \\varepsilon) D_n +\n(r - 2\\sqrt{\\varepsilon}) D_n}{2}\n\\subseteq\n\\]\n\\[\n\\frac{K + 2\\sqrt{\\varepsilon}D_n +\n(r - 2\\sqrt{\\varepsilon}) D_n}{2} =\n\\frac{K + r  D_n}{2} \n\\subseteq\nM_u (K).\n\\]\nSo the difference $(1-[2\\sqrt{\\varepsilon} + \\varepsilon])- r$\ndecreases by at least half. The inner radius exceeds $1 -\n4\\sqrt{\\varepsilon}$ if and only if its difference from \n$(1-[2\\sqrt{\\varepsilon} + \\varepsilon])$ decreases below\n$2\\sqrt{\\varepsilon} - \\varepsilon > \\sqrt{\\varepsilon}$. Thus it\nsuffices to decrease that difference to $\\sqrt{\\varepsilon}$. For that\nwe require no more than $N_b=|\\log_2(\\sqrt{\\varepsilon})| = \n\\frac{1}{\\log 4} |\\log(\\varepsilon)|$ symmetrizations.\n\\end{proof}\n\n\\begin{proof}[{\\bf Proof of Theorem \\ref{Thm_Mink-Isometric}.}] To\ncomplete the proof one has to combine the steps above. Let $K\\in\\mathscr{S}^n$,\nsuch that $M^\\ast(K) = 1$. By Lemma \\ref{Lem_Give-Small-Ball}, there\nexist $cn$ Minkowski symmetrizations that transform $K$ into a set\n$K_1$ such that\n\\[\n\\frac{c_2}{\\sqrt{n}}D_n \\subset K_1\n\\]\nand no Minkowski symmetrization can change this fact. Let $0 <\n\\varepsilon < 1/25$. By Theorem \\ref{thm-Klartag}, there exist\n$cn|\\log\\varepsilon|$ symmetrization that transform ${\\rm conv}(K_1)$ into\na convex set $L$ such that $(1-\\varepsilon)D_n \\subset L \\subset (1 +\n\\varepsilon)D_n$. Since $M_u {\\rm conv} K = {\\rm conv} M_u K$, we may apply the\nsame symmetrizations to $K_1$ to obtain a new set $K_2$ such that\n${\\rm conv}(K_2) = L$. By Lemma \\ref{Lem_Conv=Ball-to-eps-net}, we get the\nfollowing:\n\\[\n(1-\\varepsilon)S^{n-1} \\subset K_2 + 2\\sqrt{\\varepsilon}D_n.\n\\]\nIn addition, \n\\[\n\\frac{c_2}{\\sqrt{n}}D_n \\subset K_2.\n\\]\nThus, by Lemma \\ref{lem-StarShapedContainsBall} there exist $\\alpha +\n\\beta|\\log\\varepsilon| + \\gamma\\log n$ Minkowski symmetrizations that\ntransform $K_2$ into $K_3$ such that\n\\[\n(1-4\\sqrt{\\varepsilon})D_n \\subset K_3.\n\\]\nRecall that $K_3 \\subset (1+\\varepsilon)D_n$. To sum it up, we\napplied no more than\n\\[\ncn + cn|\\log\\varepsilon| + \\alpha + \\beta|\\log\\varepsilon| +\n\\gamma\\log n \\le C n |\\log\\varepsilon|\n\\]\nsymmetrizations, for some universal constant $C > 0$.\n\nDuring the proof we assumed that $\\varepsilon < 1/25$. This can be\neasily changed to $\\varepsilon < 1/2$, at the cost of a different\nconstant in the expression $C n |\\log\\varepsilon|$, by always\nsymmetrizing the set to be $\\varepsilon / 25$ close to the Euclidean\nball. By the same argument one may extend for all $\\varepsilon\\in\n(0,1)$, and the corresponding bound on the number of symmetrizations\nwill become $n (C|\\log\\varepsilon| + C')$.\n\\end{proof}\n\n\n", "meta": {"timestamp": "2015-05-18T02:11:08", "yymm": "1505", "arxiv_id": "1505.04151", "language": "en", "url": "https://arxiv.org/abs/1505.04151", "abstract": "We provide sharp upper bounds for the number of symmetrizations required to transform a star shaped set in ${\\mathbb R}^n$ arbitrarily close (in the Hausdorff metric) to the Euclidean ball.", "subjects": "Metric Geometry (math.MG); Functional Analysis (math.FA)", "title": "Minkowski Symmetrizations of Star Shaped Sets", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9899864299677055, "lm_q2_score": 0.8499711775577735, "lm_q1q2_score": 0.8414599316458669}}
{"url": "https://arxiv.org/abs/1606.03918", "title": "Error Analysis of Lagrange Interpolation on Tetrahedrons", "abstract": "This paper describes the analysis of Lagrange interpolation errors on tetrahedrons. In many textbooks, the error analysis of Lagrange interpolation is conducted under geometric assumptions such as shape regularity or the (generalized) maximum angle condition. In this paper, we present a new estimation in which the error is bounded in terms of the diameter and projected circumradius of the tetrahedron. Because we do not impose any geometric restrictions on the tetrahedron itself, our error estimation may be applied to any tetrahedralizations of domains including very thin tetrahedrons.", "text": "\\section{Introduction}\\label{intro}\nLagrange interpolation on tetrahedrons and the associated error\nestimates are important subjects in numerical analysis. In particular,\nthey are crucial in the error analysis of finite element methods.\nThroughout this paper, $K \\subset \\mathbb R^3$ denotes a tetrahedron\nwith vertices $\\mathbf{x}_i$, $i=1,\\cdots,4$, and tetrahedrons are assumed to\nbe closed sets. Let $\\lambda_i$ be the barycentric coordinates of a\ntetrahedron with respect to $\\mathbf{x}_i$. By definition,\nwe have $0 \\le \\lambda_i \\le 1$, $\\sum_{i=1}^{4} \\lambda_i =1$.  Let $\\mathbb N_0$\nbe the set of nonnegative integers, and\n$\\gamma = (a_1,\\cdots,a_{4}) \\in \\mathbb N_0^{4}$ be a multi-index.\nLet $k$ be a positive integer. If $|\\gamma| := \\sum_{i=1}^{4}a_i = k$, then\n${\\gamma}/{k} := \\left({a_1}/{k}, \\cdots, {a_{4}}/{k}\\right)$\ncan be regarded as a barycentric coordinate in $K$.\nThe set $\\Sigma^k(K)$ of points on $K$ is defined by\n\\begin{equation*}\n   \\Sigma^k(K) := \\left\\{\n   \\frac{\\gamma}{k} \\in K \\Bigm| |\\gamma| = k, \\;\n     \\gamma \\in \\mathbb N_0^{4} \\right\\}.\n\\end{equation*}\nFor $k$, $\\mathcal{P}_k$ is the set of all polynomials of three variables\nwhose degree is at most $k$.  For a continuous function $v \\in C^0(K)$, \nthe Lagrange interpolation $\\mathcal{I}_K^k v \\in \\mathcal{P}_k$ of degree $k$ is defined as\n\\[\n   v(\\mathbf{x}) = (\\mathcal{I}_K^k v)(\\mathbf{x}), \\quad \\forall \\mathbf{x} \\in \\Sigma^k(K).\n\\]\nWe attempt to obtain an upper bound of the error\n$|v - \\mathcal{I}_K^k v|_{m,p,K}$ for integers $0 \\le m \\le k$, where\n$|\\cdot|_{m,p,K}$ is the usual Sobolev semi-norm.\n\nFor now, let $K$ be a $d$-simplicial element ($d = 2, 3$).  Let\n$h_K := \\mathrm{diam}(K)$ and $\\rho_K$ be the diameter of its inscribed\nsphere.  For the error analysis of Lagrange interpolation on simplicial\nelements, many textbooks on finite element methods, such as those by\nCiarlet \\cite{Ciarlet}, Brenner and Scott \\cite{BrennerScott}, and\nErn and Guermond \\cite{ErnGuermond}, explain the following theorem.\n\n\\begin{theorem}[Shape-regularity]\\label{thm:S-regularity}\nLet $\\sigma > 0$ be a constant.\nIf $h_K/\\rho_K \\le \\sigma$, then there\n exists a constant $C=C(\\sigma)$ independent of $h_K$ such that\n\\begin{equation}\n  |v - \\mathcal{I}_K^1 v|_{1,2,K} \\le C  h_K |v|_{2,2,K},\n   \\qquad  \\forall v \\in H^2(K).\n  \\label{eq:standard-est}\n\\end{equation}\n\\end{theorem}\nNote that \\eqref{eq:standard-est} does not hold if $d \\ge 4$.\n\nFor the case of triangles, shape-regularity is equivalent to\nthe \\textit{minimum angle condition} \\cite{Zlamal}.  It is known\nthat shape-regularity is not an optimal condition on the geometry\nof triangles.  If the maximum angle of a triangle is less than a fixed\nconstant $\\theta_1 < \\pi$, then the estimation \\eqref{eq:standard-est}\nholds with a constant $C = C(\\theta_1)$.  This condition is known as\nthe \\textit{maximum angle condition} \\cite{BabuskaAziz, Jamet}. \n\nThe present authors recently reported\nan error estimation in terms of the circumradius of a triangle\n\\cite{KobayashiTsuchiya1, KobayashiTsuchiya3, KobayashiTsuchiya4}.\nLet $R_K$ be the circumradius of a triangle $K$.\n\\begin{theorem}[Circumradius estimates]\\label{thm:circumradius-est}\nLet $K$ be an arbitrary triangle.  Let\n$1 \\le p \\le \\infty$, and $k$, $m$ be integers such that $k \\ge 1$ and\n$0 \\le  m \\le k$.  Then, for the $k$th-order Lagrange interpolation\n $\\mathcal{I}_K^k$ on $K$, the estimation\n\\begin{equation} \\label{eq:circumradius-est}\n   |v - \\mathcal{I}_K^k v|_{m,p,K} \\le C\n   \\left(\\frac{R_K}{h_K}\\right)^m h_K^{k+1-m} |v|_{k+1,p,K}\n  =  C R_K^m h_K^{k+1-2m} |v|_{k+1,p,K}\n\\end{equation}\nholds for any $v \\in W^{k+1,p}(K)$,\nwhere the constant $C=C(k,m,p)$ is independent of the geometry of $K$.\n\\end{theorem}\nNote that the circumradius estimation \n\\eqref{eq:circumradius-est} is closely related to the definition of\nsurface area \\cite{KobayashiTsuchiya2}.\n\nThe aim of this paper is to derive a similar error estimation\nto Theorem~\\ref{thm:circumradius-est} for tetrahedrons under no specific\ngeometric restrictions.\n\nTo extend the circumradius estimation \\eqref{eq:circumradius-est}\nto tetrahedrons, an immediate idea is to replace the circumradius\nof a triangle with the radius of the circumsphere of a tetrahedron.\nHowever, this idea can be immediately rejected by considering \nthe tetrahedron $K$ with vertices\n$\\mathbf{x}_1 = (h,0,0)^\\top$, $\\mathbf{x}_2 = (-h,0,0)^\\top$,\n$\\mathbf{x}_3 = (0,-h,h^\\alpha)^\\top$, and $\\mathbf{x}_4 = (0,h,h^\\alpha)^\\top$\nwith $h > 0$ and $\\alpha > 0$.\nSetting $v_1(x,y,z) := x^2 - h^2 + h^{2-\\alpha}z$,\nwe have that $\\mathcal{I}_K^1 v_1 \\equiv 0$, and a simple computation yields\n$|v_1 - \\mathcal{I}_K^1 v_1|_{1,\\infty,K} = |v_1|_{1,\\infty,K} \\ge h^{2-\\alpha}$\nand $|v_1|_{2,\\infty,K} = 2$.  Hence, if $\\alpha > 2$,\nan inequality such as \\eqref{eq:circumradius-est} does\nnot hold for tetrahedrons, although the radius of the circumsphere of the above $K$\nconverges to $0$ as $h \\to 0$.  Tetrahedrons such as the above $K$ are\ncalled \\textit{slivers} \\cite{Sliver}.\n\nThus, we introduce the projected circumradius, also denoted by\n$R_K$, of a tetrahedron $K$ in Section~\\ref{section:Circumradius}.\nWe then obtain an error estimation for tetrahedrons that is\nfundamentally similar to \\eqref{eq:circumradius-est}\nin Theorem~\\ref{thm:main}.\n\nIn Section~\\ref{Sect:Prei}, we state some preliminary results used in this paper.\nIn particular, we define the standard position and the projected\ncircumradius for tetrahedrons.  In Section~\\ref{Sect:Squeezing}, we\nrecall definitions related to the quotient differences of functions\nwith multi-variables. We then reconfirm the Squeezing Theorem for\ntetrahedrons.  In Section~\\ref{sect:singular}, we obtain the error\nestimation of Lagrange interpolation in terms of the singular values of\na linear transformation.  In Section~\\ref{sec:geometric}, we present\na geometric interpretation of the singular values of the linear\ntransformation, and finally obtain the main theorem.\n\n\n\\section{Preliminaries}\\label{Sect:Prei}\n\\subsection{Notation}\nLet $d \\ge 1$ be a positive integer and $\\mathbb R^d$ be the $d$-dimensional\nEuclidean space.  We denote the Euclidean norm of\n$\\mathbf{x} = (x_1,\\cdots,x_d)^\\top \\in \\mathbb R^d$ by $|\\mathbf{x}|$. \nWe always regard $\\mathbf{x} \\in \\mathbb R^d$ as a column vector.\nFor a matrix $A$ and $\\mathbf{x} \\in \\mathbb R^d$, $A^\\top$ and $\\mathbf{x}^\\top$\ndenote their transpositions.\n\nFor $\\delta = (\\delta_1,...,\\delta_d) \\in \\mathbb N_{0}^d$,\nthe multi-index $\\partial^\\delta$ of partial differentiation \n(in the sense of the distribution) is defined by\n\\[\n    \\partial^\\delta = \\partial_\\mathbf{x}^\\delta\n    := \\frac{\\partial^{|\\delta|}\\ }\n   {\\partial x_1^{\\delta_1}...\\partial x_d^{\\delta_d}}, \\qquad\n   |\\delta| := \\delta_1 + ... + \\delta_d.\n\\]\n\nLet $\\Omega \\subset \\mathbb R^d$ be a (bounded) domain.  The usual Lebesgue\nspace is denoted by $L^p(\\Omega)$ for $1 \\le p \\le \\infty$.\nFor a positive integer $k$, the Sobolev space $W^{k,p}(\\Omega)$ is\ndefined by\n$\\displaystyle\n  W^{k,p}(\\Omega) := \n  \\left\\{v \\in L^p(\\Omega) \\, | \\, \\partial^\\delta v \\in L^p(\\Omega), \\,\n   |\\delta| \\le k\\right\\}$.\nFor $1 \\le p < \\infty$, the norm and semi-norm of $W^{k,p}(\\Omega)$ are defined\nby\n\\begin{gather*}\n  |v|_{k,p,\\Omega} := \n  \\biggl(\\sum_{|\\delta|=k} |\\partial^\\delta v|_{0,p,\\Omega}^p\n   \\biggr)^{1/p}, \\quad   \\|v\\|_{k,p,\\Omega} := \n  \\biggl(\\sum_{0 \\le m \\le k} |v|_{m,p,\\Omega}^p \\biggr)^{1/p}\n\\end{gather*}\nand $\\displaystyle   |v|_{k,\\infty,\\Omega} := \n  \\max_{|\\delta|=k} \\left\\{\\mathrm{ess}\n   \\sup_{\\hspace{-5mm}\\mathbf{x} \\in\\Omega}|\\partial^\\delta v(\\mathbf{x})|\\right\\}$,\n $\\displaystyle   \\|v\\|_{k,\\infty,\\Omega} := \n  \\max_{0 \\le m  \\le k} \\left\\{|v|_{m,\\infty,\\Omega}\\right\\}$.\n\n\n\\subsection{The imbedding theorem}\nLet $1 < p \\le \\infty$. From Sobolev's imbedding theorem and Morry's\ninequality, we have the continuous imbeddings\n\\begin{gather*}\n   W^{2,p}(K) \\subset C^{1,1-3/p}(K), \\quad p > 3, \\\\\n   W^{2,3}(K) \\subset W^{1,q}(K) \\subset C^{0,1-3/q}(K),\n      \\quad \\forall q > 3, \\\\\n  W^{2,p}(K) \\subset W^{1,3p/(3-p)}(K) \\subset C^{0,2-3/p}(K),\n      \\quad \\frac{3}{2} < p < 3, \\\\\n  W^{3,3/2}(K) \\subset W^{2,3}(K) \\subset W^{1,q}(K)\n  \\subset C^{0,1-3/q}(K), \\quad \\forall q > 3, \\\\\n  W^{3,p}(K) \\subset W^{2,3p/(3-p)}(K) \\subset W^{1,3p/(3-2p)}(K)\n  \\subset C^{0,3-3/p}(K), \\quad 1 < p < \\frac{3}{2}.\n\\end{gather*}\nAlthough Morry's inequality may not be applied, the continuous\nimbedding $W^{3,1}(K)$ $\\subset C^{0}(K)$ still holds.\nFor the imbedding theorem, see \\cite{AdamsFournier},\n\\cite{Brezis}, and \\cite{KJF}.  In the following, we assume that $p$\nis such that the imbedding $W^{k+1,p}(K) \\subset C^{0}(K)$\nholds, that is,\n\\begin{align*}\n     1 \\le p \\le \\infty, \\quad  \\text{ if } k+1 \\ge 3\n      \\quad \\text{ and } \\quad\n    \\frac{3}{2} < p \\le \\infty, \\quad \\text{ if } k+1 =2.\n\\end{align*}\n\n\\subsection{The reference tetrahedrons}\nLet $\\widehat{K}$ and $\\widetilde{K}$ be tetrahedrons that have the \nfollowing vertices (see Figure~\\ref{ref_tetra}):\n\\begin{align*}\n  \\widehat{K} \\text{ has the vertices } \\; (0,0,0)^\\top, \\; (1,0,0)^\\top, \\;\n  (0,1,0)^\\top, \\; (0,0,1)^\\top, \\\\\n  \\widetilde{K} \\text{ has the vertices } \\; (0,0,0)^\\top, \\; (1,0,0)^\\top, \\;\n  (1,1,0)^\\top, \\; (0,0,1)^\\top.\n\\end{align*}\n\\begin{figure}[tbhp]\n\\centering\n\\includegraphics[width=9cm]{reference-tetra-crop.eps}\n\\caption{The reference tetrahedrons $\\widehat{K}$ and $\\widetilde{K}$.}\n\\label{ref_tetra}\n\\end{figure}\nThese tetrahedrons are called the \\textbf{reference tetrahedrons}.\nIn this paper, we denote the reference tetrahedrons by \n$\\mathbf{K}$, that is, $\\mathbf{K}$ is either of $\\{\\widehat{K},\\widetilde{K}\\}$.\n\n\n\\subsection{Standard position of tetrahedrons}\nWhen considering tetrahedrons, it is convenient to define their ``standard\ncoordinates.''  Take any tetrahedron $K$ with vertices\n$\\mathbf{x}_i$, $i = 1, \\dots, 4$.  The facet $B$ with vertices $\\mathbf{x}_1$,\n$\\mathbf{x}_2$, $\\mathbf{x}_3$ is regarded as the base of $K$. Let $\\alpha$ and\n$\\beta$, $0 < \\beta \\le \\alpha$, be the longest and shortest\nlengths of the edges of $B$.   Without loss of generality, we assume\nthat $\\mathbf{x}_1\\mathbf{x}_2$ is the longest edge of $B$;\n$|\\mathbf{x}_1 - \\mathbf{x}_2| = \\alpha$.     Consider cutting $\\mathbb R^3$\nwith the plane that contains the midpoint of the edge $\\mathbf{x}_1\\mathbf{x}_2$ and\nis perpendicular to the vector $\\mathbf{x}_1 - \\mathbf{x}_2$.  Then, there exist\ntwo cases: (i) $\\mathbf{x}_3$ and $\\mathbf{x}_4$ belong to the same half-space,\nor (ii) $\\mathbf{x}_3$ and $\\mathbf{x}_4$ belong to different half-spaces.\nLet $\\gamma := |\\mathbf{x}_1 - \\mathbf{x}_4|$.\nUnder appropriate rotation, translation, and reflection operations,\nthese situations can be written using the parameters\n\\begin{equation}\n \\begin{cases}\n  0 < \\beta \\le \\alpha, \\quad 0 < \\gamma, \\quad s_1^2 + t_1^2  = 1, \\;\n s_1 > 0, \\; t_1 > 0, \\quad \\beta s_1 \\le \\frac{\\alpha}{2}, \\\\\n s_{21}^2 + s_{22}^2 + t_2^2 = 1, \\; t_2 > 0, \\quad \n  \\gamma s_{21} \\le \\frac{\\alpha}{2},\n \\end{cases}\n \\label{eq:tetra-param}\n\\end{equation}\nas\n\\begin{subequations} \\label{eq:position}\n\\begin{gather}\n \\mathbf{x}_1 = (0,0,0)^\\top, \\; \\mathbf{x}_2 = (\\alpha,0,0)^\\top, \\;\n  \\mathbf{x}_4 = (\\gamma s_{21}, \\gamma s_{22}, \\gamma t_2)^\\top,\\\\\n  \\begin{cases}\n   \\mathbf{x}_3 = (\\beta s_1, \\beta t_1,0)^\\top & \\text{for the case (i)} \\\\\n   \\mathbf{x}_3 = (\\alpha - \\beta s_1, \\beta t_1,0)^\\top & \\text{for the case (ii)}\n  \\end{cases}.\n\\end{gather}\n\\end{subequations}\nWe refer to the coordinates in \\eqref{eq:position}\nas the \\textit{standard position} of a tetrahedron.  In the following,\nwe sometimes write $h_B := \\alpha$.  Let $R_B$ be the circumradius\nof $B$.\n\n\n\\subsection{The projected circurmradius of tetrahedrons}\n\\label{section:Circumradius}\nSuppose that a tetrahedron $K$ is at the standard position.\nLet $\\theta \\in \\mathbb R$ be such that\n$-\\frac{\\pi}{2} \\le \\theta \\le \\frac{\\pi}{2}$.\nLet $\\delta_\\theta$ be the linear transformation defined by\nthe matrix\n\\[\n   \\begin{pmatrix}\n      \\cos\\theta & -\\sin\\theta & 0 \\\\\n          0 & 0 & 0 \\\\\n          0 & 0 & 1\n   \\end{pmatrix}\n   = \\begin{pmatrix}\n      1 & 0 & 0 \\\\\n      0 & 0 & 0 \\\\\n      0 & 0 & 1\n   \\end{pmatrix}\n   \\begin{pmatrix}\n      \\cos\\theta & -\\sin\\theta & 0 \\\\\n      \\sin\\theta & \\cos\\theta & 0 \\\\\n          0 & 0 & 1\n   \\end{pmatrix}.\n\\]\nThat is, $\\delta_\\theta$ is a composite transformation of rotation\nabout the $z$-axis with angle $\\theta$ and projection to the $xz$-plane.\nThe image $\\delta_\\theta(K)$ is a triangle on the $xz$-plane.  Let\n$R_\\theta$ be the circumradius of $\\delta_\\theta(K)$.  Define\n\\[\n   R_P := \\max_{\\theta \\in [-\\pi/2,\\pi/2]} R_\\theta.\n\\]\nThe \\textit{projected circumradius} $R_K$ of a tetrahedron $K$ is\ndefined by\n\\begin{equation}\\label{def:RK}\n   R_K := \\min_{B} \\frac{R_BR_P}{h_B},\n\\end{equation}\nwhere the minimum is taken over all the facets of $K$.\n\n\\subsection{The setting of error estimation}\nWe define the set $\\mathcal{T}_p^{k}(K) \\subset W^{k+1,p}(K)$ by\n\\begin{gather*}\n   \\mathcal{T}_p^{k}(K) := \\left\\{v \\in W^{k+1,p}(K) \\Bigm|\n     v(\\mathbf{x}) = 0, \\; \\forall \\mathbf{x} \\in \\Sigma^k(K) \\right\\}.\n\\end{gather*}\nFor Lagrange interpolation $\\mathcal{I}_K^k(v)$, it is clear from the definition that\n\\[\n     v - \\mathcal{I}_K^k v \\in \\mathcal{T}_p^{k}(K), \\quad \\forall \n     v \\in W^{k+1,p}(K).\n\\]\nFor an integer $m$ such that $0 \\le m \\le k$, $B_p^{m,k}(K)$\nis defined by\n\\begin{gather*}\n    B_p^{m,k}(K) := \\sup_{v \\in \\mathcal{T}_p^{k}(K)}\n    \\frac{|v|_{m,p,K}}{|v|_{k+1,p,K}}.\n\\end{gather*}\nNote that we have\n\\begin{gather*}\n B_p^{m,k}(K) = \\inf\\left\\{C ; \n   |v - \\mathcal{I}_K^k v|_{m,p,K} \\le C |v|_{k+1,p,K}, \\\n   \\forall v \\in W^{k+1,p}(K)\\right\\},\n\\end{gather*}\nthat is, $B_p^{m,k}(K)$ is the \\textit{best} constant $C$ for the\nerror estimation\n\\[\n   |v - \\mathcal{I}_K^k v|_{m,p,K} \\le C |v|_{k+1,p,K}, \\qquad \n    \\forall v \\in W^{k+1,p}(K).\n\\]\nTherefore, we try to obtain an upper bound of $B_p^{m,k}(K)$ in terms of\ngeometric quantities of $K$.\n\n\n\\section{The Squeezing Theorem}\\label{Sect:Squeezing}\nLet $a$, $b \\in\\mathbb R$ be such that $0 < a \\le 1$ and $0 < b$.\nWe then define the \\textit{squeezing map} $sq_1^{ab}:R^3 \\to \\mathbb R^3$ by\n\\[\n  sq_1^{ab}(x,y,z) := (x, a y, b z)^\\top,\n   \\quad (x,y,z)^\\top \\in \\mathbb R^3.\n\\]\nLet $K_{ab} := sq_1^{ab}(\\mathbf{K})$.  In this section, we will prove the following theorem.\n\n\\begin{theorem}\\label{squeezingtheorem}\nLet $k$ be a positive integer and $1 \\le p \\le \\infty$.  There exists a\nconstant $C_{k,m,p}$ independent of $a$ and $b$ such that, for $m = 0, \\cdots, k$,\n\\begin{gather*}\n   B_p^{m,k}(K_{ab}) := \\hspace{-0.2cm}\n   \\sup_{v \\in \\mathcal{T}_p^{k}(K_{ab})}\n  \\frac{|v|_{m,p,K_{ab}}}{|v|_{k+1,p,K_{ab}}}\n   \\le \\max\\{1,b^{k+1-m}\\}C_{k,m,p},\n\\end{gather*}\nwhere $p$ is taken such that\n\\begin{gather}\n  \\begin{cases}\n      2 < p \\le \\infty & \\text{ if } k - m = 0, \\\\ \n      \\frac{3}{2} < p \\le \\infty & \\text{ if } k = 1, \\; m = 0,\\\\\n      1 \\le p \\le \\infty & \\text{ if } k \\ge 2 \\text{ and } \\; k-m \\ge 1.\n   \\end{cases}\n   \\label{p-cond}\n\\end{gather}\n\\end{theorem}\nAlthough the proof of Theorem~\\ref{squeezingtheorem} is very similar to that of\n\\cite[Theorem~1.3]{KobayashiTsuchiya4}, we provide a sketch here for\nreaders' convenience.  Note that the restriction \n$2 < p \\le \\infty$ for the case $k=m$ comes from the continuity of\nthe trace operator $t:W^{1,p}(\\mathbf{K}) \\ni v \\mapsto v|_S \\in L^1(S)$, where\n$S \\subset \\mathbf{K}$ is a non-degenerate segment (see\n\\cite[Section~3]{KobayashiTsuchiya4}).\nUsing the counter-example given by Shenk \\cite{Shenk}, we find that\nthis restriction cannot be improved.\n\nTo prove Theorem~\\ref{squeezingtheorem}, we recall the definitions of\ndifference quotients of multi-variable functions, the rectangular\nparallelepiped $\\square_\\gamma^\\delta$ defined by the lattice points\n$\\mathbf{x}_\\gamma$, $\\Delta^\\delta\\mathbf{x}_\\gamma$ in $\\mathbf{K}$, and the integral\n$\\int_{\\square_\\gamma^\\delta} v$ introduced in\n\\cite[Section~2]{KobayashiTsuchiya4}.  The set\n$\\Xi_p^{\\delta,k} \\subset W^{k+1-|\\delta|,p}(\\mathbf{K})$ is then defined by\n\\begin{align*}\n  \\Xi_p^{\\delta,k} & := \\left\\{ v \\in W^{k+1-|\\delta|,p}(\\mathbf{K}) \\Bigm| \n    \\int_{\\square_{\\gamma}^{\\delta}} v = 0,\\quad \\forall\n   \\square_{\\gamma}^{\\delta} \\subset \\mathbf{K} \\right\\}.\n\\end{align*}\nNote that $u \\in \\mathcal{T}_p^{k}(\\mathbf{K})$ implies\n$\\partial^\\delta u \\in \\Xi_p^{\\delta,k}$.\n\n\\begin{lemma}\\label{lem32} \nWe have $\\Xi_p^{\\delta,k} \\cap \\mathcal{P}_{k-|\\delta|} = \\{0\\}$.\nThat is, if $q \\in \\mathcal{P}_{k-|\\delta|}$ belongs to $\\Xi_p^{\\delta,k}$,\nthen $q=0$.\n\\end{lemma}\n\n\\textit{Proof.}\n Note that\n$\\mathrm{dim} \\mathcal{P}_{k-|\\delta|} = \\#\\{\\square_{lp}^\\delta \\subset \\mathbf{K} \\}$.\nFor example, if $k=4$ and $|\\delta|=3$, then\n$\\mathrm{dim}\\mathcal{P}_1 = 4$.  This corresponds to the fact that,\nin $\\mathbf{K}$, there are four cubes of size $1/4$  for $\\delta=(1,1,1)$ and\nthere are four rectangles of size $1/2 \\times 1/4$ for $\\delta=(1,2,0)$.\nAll their vertices (corners) belong to $\\Sigma^4(\\mathbf{K})$\n(see Figure~2). Now, suppose that $v \\in \\mathcal{P}_{k-|\\delta|}$ satisfies\n$\\int_{\\square_{lp}^{\\delta}} q = 0$ for all\n$\\square_{lp}^{\\delta} \\subset \\mathbf{K}$.  These conditions are linearly\nindependent and determine $q = 0$ uniquely. \\hfill $\\square$\n\n\\begin{figure}[htbp]\\label{cubes_tetra}\n\\centering\n\\subfloat{\\includegraphics[width=4.5cm]{kyoko.eps}} \\qquad\n\\subfloat{\\includegraphics[width=3cm]{takako.eps}} \\qquad\n\\subfloat{\\includegraphics[width=3cm]{michiko.eps}}\n\\caption{The four cubes and four rectangles in $\\mathbf{K}$.}\n\\end{figure}\n\n\\vspace{0.3cm}\nThe constant $A_p^{\\delta,k}$ is defined by\n\\begin{align*}\n   A_p^{\\delta,k} := \\sup_{v \\in \\Xi_p^{\\delta,k}} \\frac{|v|_{0,p,\\mathbf{K}}}\n  {|v|_{k+1-|\\delta|,p,\\mathbf{K}}}.\n\\end{align*}\nThe following lemma is an extension of \\cite[Lemma~2.1]{BabuskaAziz}.\n\n\\begin{lemma}\\label{lem33}\nLet $p$ be such that $2 < p \\le \\infty$ if $k+1-|\\delta|=1$ or\n$1 \\le p \\le \\infty$ if $k+1-|\\delta|\\ge 2$.  We then have\n$A_p^{\\delta,k} < \\infty$.\n\\end{lemma}\n\n\\textit{Proof.}\nSee the proof of \\cite[Lemma~3.3]{KobayashiTsuchiya4}.\n\\hfill $\\square$\n\n\\vspace{0.3cm}\n\\textit{Proof of} Theorem~\\ref{squeezingtheorem}.\nFirst, let $1 \\le p < \\infty$ and $1 \\le m \\le k$.\nFor a multi-index $\\gamma = (n_1, n_2, n_3) \\in \\mathbb N_0^3$ and a real $t \\neq 0$,\nset $(a, b)^{\\gamma t} := a^{n_2 t}b^{n_3 t}$.  Take an arbitrary \n$v \\in \\mathcal{T}_p^k(K_{ab})$ and pull it back to\n$u := v\\circ sq_1^{ab} \\in \\mathcal{T}_p^k(\\mathbf{K})$.   We then have\n{\\allowdisplaybreaks\n\\begin{align*}\n  \\frac{|v|_{m,p,K_{ab}}^p}{|v|_{k+1,p,K_{ab}}^p}\n   & = \\frac{\\sum_{|\\gamma| = m} \\frac{m!}{\\gamma!}\n      (a,b)^{-\\gamma p}\\left|\\partial^{\\gamma}u\\right|_{0,p,\\mathbf{K}}^p}\n      {\\sum_{|\\delta|= k+1} \\frac{(k+1)!}{\\delta!}\n     (a,b)^{- \\delta p} \\left|\\partial^{\\delta} u \\right|_{0,p,\\mathbf{K}}^p} \\\\\n  & = \\frac{\\sum_{|\\gamma| = m}\\frac{m!}{\\gamma!}(a,b)^{-\\gamma p}\n         \\left|\\partial^{\\gamma}u \\right|_{0,p,\\mathbf{K}}^p }\n       {\\sum_{|\\gamma| = m} \\frac{m!}{\\gamma!}(a,b)^{-\\gamma p}\n       \\left(\\sum_{|\\eta| = k+1-m} \\frac{(k+1-m)!}\n             {\\eta!(a,b)^{\\eta p} }\\left|\\partial^{\\eta}\n                 (\\partial^{\\gamma}u)\\right|_{0,p,\\mathbf{K}}^p\\right)} \\\\\n  & \\le \\frac{\\max\\{1,b^{(k+1-m)p}\\}\n   \\sum_{|\\gamma| = m}\\frac{m!}{\\gamma!}(a,b)^{-\\gamma p}\n         \\left|\\partial^{\\gamma}u \\right|_{0,p,\\mathbf{K}}^p }\n       {\\sum_{|\\gamma| = m} \\frac{m!}{\\gamma!}(a,b)^{-\\gamma p}\n       \\left(\\sum_{|\\eta| = k+1-m} \\frac{(k+1-m)!}\n             {\\eta! }\\left|\\partial^{\\eta}\n                 (\\partial^{\\gamma}u)\\right|_{0,p,\\mathbf{K}}^p\\right)} \\\\\n   & = \\max\\{1,b^{(k+1-m)p}\\}\n\\frac{\\sum_{|\\gamma| = m} \\frac{m!}{\\gamma!}(a,b)^{-\\gamma p}\n         \\left|\\partial^{\\gamma}u \\right|_{0,p,\\mathbf{K}}^p}\n       {\\sum_{|\\gamma| = m} \\frac{m!}{\\gamma!}\n      (a,b)^{-\\gamma p}\n        \\left|\\partial^{\\gamma}u \\right|_{k+1-m,p,\\mathbf{K}}^p} \\\\\n    & \\le \\max\\{1,b^{(k+1-m)p}\\}\\max_{|\\gamma|=m} \n    \\left(A_p^{\\gamma,k}\\right)^p\n   = \\left(\\max\\{1,b^{k+1-m}\\}C_{k,m,p}\\right)^p,\n\\end{align*}\n}\nwhere $C_{k,m,p} := \\max_{|\\gamma|=m} A_p^{\\gamma,k}$.\nProofs for the cases $1 \\le p < \\infty$ with $m=0$ and\n$p = \\infty$ with $0 \\le m \\le k$ are very similar.  See\nthe proof of \\cite[Theorem~1.3]{KobayashiTsuchiya4}.  \n\\hfill $\\square$\n\n\n\\vspace{0.3cm}\nWe now generalize the squeezing map.\nLet $\\alpha$, $\\beta$, and $\\gamma \\in\\mathbb R$ be such that\n$0 < \\beta \\le \\alpha$ and $0 < \\gamma$.\nWe then define the \\textit{squeezing map}\n$sq_2^{\\alpha\\beta\\gamma}:\\mathbb R^3 \\to \\mathbb R^3$ by\n\\[\n  sq_2^{\\alpha\\beta\\gamma}(x,y,z) := (\\alpha x, \\beta y, \\gamma z)^\\top,\n   \\quad (x,y,z)^\\top \\in \\mathbb R^3.\n\\]\nLet $K_{\\alpha\\beta\\gamma} := sq_2^{\\alpha\\beta\\gamma}(\\mathbf{K})$.  Note that\n$G_{\\alpha} := sq_2^{\\alpha\\alpha\\alpha}$ is a similar transformation.\nLet $\\Omega \\subset \\mathbb R^3$ be a domain and an arbitrary function \n$v \\in W^{k,p}(\\Omega)$ be pulled-back to\n$u := v \\circ G_\\alpha \\in W^{k,p}(\\Omega_\\alpha)$ with\n$\\Omega_\\alpha := G_{1/\\alpha}(\\Omega)$.\nIt is straightforward to check that \n\\[\n  |v|_{k,p,\\Omega} = \\alpha^{3/p-k}|u|_{k,p,\\Omega_\\alpha}.\n\\]\n\nNow, take an arbitrary function $v \\in \\mathcal{T}_p^k(K_{\\alpha\\beta\\gamma})$ and define\n$u := v\\circ sq_2^{\\alpha\\beta\\gamma}$.  Then,\n$u = u_1 \\circ sq_1^{\\frac{\\beta}{\\alpha},\\frac{\\gamma}{\\alpha}}$ with\n$u_1 := v \\circ G_\\alpha$, because\n$sq_2^{\\alpha\\beta\\gamma} = G_\\alpha\\circ sq_1^{\\frac{\\beta}{\\alpha},\\frac{\\gamma}{\\alpha}}$.\nTherefore, it follows from Theorem~\\ref{squeezingtheorem} that\n\\begin{align*}\n   \\frac{|v|_{m,p,K_{\\alpha\\beta\\gamma}}}{|v|_{k+1,p,K_{\\alpha\\beta\\gamma}}} = \\alpha^{k+1-m}\n    \\frac{|u_1|_{m,p,K_{\\frac{\\beta}{\\alpha},\\frac{\\gamma}{\\alpha}}}}\n   {|u_1|_{k+1,p,K_{\\frac{\\beta}{\\alpha},\\frac{\\gamma}{\\alpha}}}}\n  & \\le \\alpha^{k+1-m} \\max \\left\\{1,\n   \\left(\\frac{\\gamma}{\\alpha}\\right)^{k+1-m}\\right\\} C_{k,m,p} \\\\\n  & = \\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-m}C_{k,m,p}.\n\\end{align*}\nHence, we have derived the following corollary.\n\n\\begin{corollary}\\label{cor:squeezing}\n Let $\\alpha$, $\\beta$, and $\\gamma \\in\\mathbb R$ be such that\n$0 < \\beta \\le \\alpha$ and $0 < \\gamma$.  \nLet $K_{\\alpha\\beta\\gamma} := sq_2^{\\alpha\\beta\\gamma}(\\mathbf{K})$. \nAssume that $k \\ge 1$, $0 \\le m \\le k$, and $p$ is taken as\n\\eqref{p-cond}.   We then have\n\\begin{align*}\n  B_p^{m,k}(K_{\\alpha\\beta\\gamma}) := \\hspace{-0.2cm} \\sup_{v \\in \\mathcal{T}_p^{k}(K_{\\alpha\\beta\\gamma})}\n  \\frac{|v|_{m,p,K_{\\alpha\\beta\\gamma}}}{|v|_{k+1,p,K_{\\alpha\\beta\\gamma}}}\n  \\le \\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-m}C_{k,m,p}. \\label{extension-3d}\n\\end{align*}\n\\end{corollary}\n\n\n\\section{Error estimates of Lagrange interpolation on general\n tetrahedrons}\n\\label{sect:singular}\nIn this section, we obtain an error estimation for Lagrange interpolation\non general tetrahedrons.  To this end, we apply the method developed in\n\\cite{KobayashiTsuchiya3}.\n\nRecall that an arbitrary tetrahedron $K$ is written as\n\\eqref{eq:position} with parameters \\eqref{eq:tetra-param}.  First, we\nconfirm that $K$ is obtained from the reference tetrahedron $\\mathbf{K}$ by an\naffine linear transformation.  Define the matrices $\\widehat{A}$,\n$\\widetilde{A}$, $G \\in GL(3,\\mathbb R)$ by\n\\begin{align*}\n   \\widehat{A} := \\begin{pmatrix}\n        1  & s_{1} & s_{21} \\\\\n        0  & t_{1} & s_{22} \\\\\n        0  & 0     & t_2    \n       \\end{pmatrix}, \\quad\n   \\widetilde{A} := \\begin{pmatrix}\n        1  & - s_{1} & s_{21} \\\\\n        0  & t_{1} & s_{22} \\\\\n        0  & 0     & t_2    \n       \\end{pmatrix}, \\quad\n  G := \\begin{pmatrix}\n        \\alpha  & 0 & 0 \\\\\n        0  & \\beta & 0 \\\\\n        0  & 0     & \\gamma    \n       \\end{pmatrix}.\n\\end{align*}\nWe then have $K = \\widehat{A}G(\\widehat{K})$ for case (i) or\n$K = \\widetilde{A}G(\\widetilde{K})$ for case (ii), that is,\n$K = \\widehat{A}(K_{\\alpha\\beta\\gamma})$ or $K = \\widetilde{A}(K_{\\alpha\\beta\\gamma})$.\nNote that $\\widehat{A}$ and $\\widetilde{A}$ can be decomposed as\n$\\widehat{A} = X\\widehat{Y}$ and $\\widetilde{A} = X\\widetilde{Y}$\nwith\n\\begin{align*}\n  X := \\begin{pmatrix}\n        1 & 0 & s_{21} \\\\\n\t0 & 1 & s_{22} \\\\\n        0 & 0 & t_2\n       \\end{pmatrix}, \\qquad\n  \\widehat{Y} := \\begin{pmatrix}\n        1 & s_1 & 0 \\\\\n\t0 & t_1 & 0 \\\\\n        0 & 0 & 1\n       \\end{pmatrix}, \\qquad\n  \\widetilde{Y} := \\begin{pmatrix}\n        1 & -s_1 & 0 \\\\\n\t0 & t_1 & 0 \\\\\n        0 & 0 & 1\n       \\end{pmatrix},\n\\end{align*}\nrespectively.  We consider the singular values of $\\widehat{A}$,\n$\\widetilde{A}$, $X$, $\\widehat{Y}$, and $\\widetilde{Y}$.\nA straightforward computation yields \n\\begin{gather*}\n  \\det\\left(X^\\top X - \\mu I\\right) \n    = (1 - \\mu)\\left(\\mu^2 - 2\\mu + t_2^2\\right), \\\\\n  \\det\\left(\\widehat{Y}^\\top \\widehat{Y} - \\mu I\\right) \n  = \\det\\left(\\widetilde{Y}^\\top \\widetilde{Y} - \\mu I\\right) \n    = (1 - \\mu)\\left(\\mu^2 - 2\\mu + t_1^2\\right),\n\\end{gather*}\nthe eigenvalues of which are\n$\\mu = 1$, $1 \\pm \\sqrt{1 - t_i^2} = 1 \\pm \\mathbf{s}_i$, $i = 1, 2$, where\n$\\mathbf{s}_1 := |s_1|$ and $\\mathbf{s}_2 := \\sqrt{s_{21}^2 + s_{22}^2}$.\nTherefore, for $\\mathbf{a} \\in \\mathbb R^3$, we have\n\\begin{gather*}\n  (1 - \\mathbf{s}_2)|\\mathbf{a}|^2 \\le |X\\mathbf{a}|^2 \\le (1 + \\mathbf{s}_2)^2|\\mathbf{a}|^2, \\\\\n  (1 - \\mathbf{s}_1)|\\mathbf{a}|^2 \\le |Z\\mathbf{a}|^2 \\le (1 + \\mathbf{s}_1)|\\mathbf{a}|^2,\n  \\qquad Z = \\widehat{Y} \\text{ or } Z = \\widetilde{Y},  \\\\\n \\prod_{i=1}^2(1 - \\mathbf{s}_i)|\\mathbf{a}|^2 \\le |V\\mathbf{a}|^2 \\le \n \\prod_{i=1}^2(1 + \\mathbf{s}_i)|\\mathbf{a}|^2, \\qquad\n V = \\widehat{A} \\text{ or } V = \\widetilde{A}.\n\\end{gather*}\n\nLet $K := V(K_{\\alpha\\beta\\gamma})$, where $V = \\widehat{A}$ or $V = \\widetilde{A}$.\nA function $v\\in W^{r,p}(K)$ is pulled-back\nto a function $u \\in W^{r,p}(K_{\\alpha\\beta\\gamma})$ by $u(\\mathbf{x}) := v(V\\mathbf{x})$.\nUsing inequality (2.1) in \\cite{KobayashiTsuchiya3}, we have\n\\begin{gather*}\n  \\frac{\\prod_{i=1}^2(1-\\mathbf{s}_i)^{r}}{t_1^{2r}t_2^{2r}}\n  \\sum_{|\\gamma|=r} (\\partial_\\mathbf{x}^\\gamma u)^2\n \\le \\sum_{|\\gamma|=r} (\\partial_\\mathbf{y}^\\gamma v)^2 \\le\n\\frac{\\prod_{i=1}^2(1+\\mathbf{s}_i)^{r}}{t_1^{2r}t_2^{2r}}\n  \\sum_{|\\gamma|=r} (\\partial_\\mathbf{x}^\\gamma u)^r,\n\\end{gather*}\nwhere $\\mathbf{y} := V\\mathbf{x}$.  The fact that\n$\\mathrm{det}A = t_1t_2$ and inequalities (2.2), (2.3) in\n\\cite{KobayashiTsuchiya3} then give\n\\begin{align*}\n   |v|_{m,p,K} & \\le 3^{m\\tau(p)} \n \\frac{\\prod_{i=1}^2(1+\\mathbf{s}_i)^{m/2}}\n       {(t_1 t_2)^{m-1}}  |u|_{m,p,K_{\\alpha\\beta\\gamma}}, \\\\\n  |v|_{k+1,p,K} & \\ge 3^{-(k+1)\\tau(p)}   \n\\frac{\\prod_{i=1}^2(1-\\mathbf{s}_i)^{(k+1)/2}}{(t_1t_2)^{k}}\n  |u|_{k+1,p,K_{\\alpha\\beta\\gamma}},   \\\\\n   \\frac{|v|_{m,p,K}}{|v|_{k+1,p,K}} & \\le 3^{(k+1+m)\\tau(p)}\n   \\frac{(t_1t_2)^{k+1-m}\\prod_{i=1}^2(1+\\mathbf{s}_i)^{m/2}|u|_{m,p,K_{\\alpha\\beta\\gamma}}}\n     {\\prod_{i=1}^2(1-\\mathbf{s}_i)^{(k+1)/2}|u|_{k+1,p,K_{\\alpha\\beta\\gamma}}} \\\\\n & = 3^{(k+1+m)\\tau(p)} \\frac{\\prod_{i=1}^2(1+\\mathbf{s}_i)^{(k+1)/2}}\n  {\\prod_{i=1}^2(1-\\mathbf{s}_i)^{m/2}}\n  \\frac{|u|_{m,p,K_{\\alpha\\beta\\gamma}}}{|u|_{k+1,p,K_{\\alpha\\beta\\gamma}}} \\\\\n  & \\le C  \\frac{\\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-m}}\n  {\\prod_{i=1}^2(1-\\mathbf{s}_i)^{m/2}},\n\\end{align*}\nwhere\n\\begin{align*}\n  \\tau(p) := \\begin{cases}\n    \\frac{1}{p} - \\frac{1}{2}, & 1 \\le p \\le 2 \\\\\n    \\frac{1}{2} - \\frac{1}{p} , & 2 \\le p \\le \\infty\n      \\end{cases}, \\qquad\n  C := 3^{(k+1+m)\\tau(p)} 2^{k+1} C_{k,m,p}.\n\\end{align*}\n(See the inequalities in \\cite[p.496]{KobayashiTsuchiya3}.)\nNote that the constant $C$ depends only on $k$, $m$, and $p$.\nHence, we have derived the following theorem:\n\\begin{theorem}\\label{thmB}\nLet $K$ be an arbitrary tetrahedron with vertices given by\n\\eqref{eq:position} with the parameters in \\eqref{eq:tetra-param}.\nLet $k$, $m$ be integers with $k \\ge 1$ and $0 \\le m \\le k$.\nLet $p$ be taken as in Theorem~\\ref{squeezingtheorem} according to $k$ and $m$.\nThen, we have\n\\[\n B_p^{m,k}(K) := \\sup_{v \\in \\mathcal{T}_p^k}\\frac{|v|_{m,p,K}}{|v|_{k+1,p,K}}\n  \\le C  \\frac{\\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-m}}\n  {\\prod_{i=1}^2(1-\\mathbf{s}_i)^{m/2}},\n\\]\nwhere $C = C(k,m,p)$ is a constant independent of $K$.\n\\end{theorem}\n\n\n\\section{A geometric interpretation and the main theorem}\n\\label{sec:geometric}\nIn this section, we consider the geometric meaning of the quantity\n$\\prod_{i=1}^2(1-\\mathbf{s}_i)^{-1/2}$ that appeared in Theorem~\\ref{thmB}.\nRecall that $K$ is a tetrahedron with vertices given by\n\\eqref{eq:position} and the parameters in \\eqref{eq:tetra-param}.\nRecall also that $B$ is the base of $K$ with vertices\n$\\mathbf{x}_1$, $\\mathbf{x}_2$, $\\mathbf{x}_3$.  The circumradius $R_B$\ncan then be written as\n\\[\n   R_B = \\frac{\\sqrt{\\alpha^2 - 2\\alpha\\beta s_1 + \\beta^2}}{2t_1}.\n\\]\nBecause\n\\begin{align*}\n   \\alpha^2 - 2\\alpha\\beta s_1 + \\beta^2 & =\n  \\frac{\\alpha^2(2 - s_1^2)}{4} + \\alpha\n   \\left(\\frac{\\alpha}{2} - \\beta s_1\\right) \n   + \\left(\\frac{\\alpha s_1}{2} - \\beta\\right)^2 \\\\\n   & \\ge \\frac{\\alpha^2(2 - s_1^2)}{4} \\ge \\frac{\\alpha^2}{4},\n\\end{align*}\nwe have\n\\begin{align}\n   R_B \\ge \\frac{\\alpha}{4 t_1} = \\frac{\\alpha}{4 \\sqrt{1 - \\mathbf{s}_1^2}} \n    \\ge \\frac{h_B}{4\\sqrt{2} \\sqrt{1 - \\mathbf{s}_1}},\n   \\label{RB-est}\n\\end{align}\nwhere we have used the definition $h_B = \\alpha$.\n\nRecall that $R_P$ was defined in Section~\\ref{section:Circumradius}.\nWe will show that there exists a constant $C$ independent of $K$\nsuch that\n\\begin{align}\n    R_P \\ge C\\frac{\\max\\{\\alpha,\\gamma\\}}{\\sqrt{1 - \\mathbf{s}_2}}.\n   \\label{RP-est}\n\\end{align}\n\nLet $\\mathbf{x}_3 = (\\eta,\\xi,0)^\\top$, that is, $\\xi = \\beta t_1$ and\n$\\eta = \\beta s_1$ or $\\eta = \\alpha - \\beta s_1$.  From the\nassumption, we have $0 < \\eta < \\alpha$ and $\\xi > 0$.  Note that \n\\begin{align*}\n   \\delta_\\theta(\\mathbf{x}_1) = (0,0,0)^\\top, \\quad\n  & \\delta_\\theta(\\mathbf{x}_2) = (\\alpha\\cos\\theta,0,0)^\\top, \\\\\n   \\delta_\\theta(\\mathbf{x}_3) = (\\eta\\cos\\theta-\\xi\\sin\\theta,0,0)^\\top, \n  \\quad &\n   \\delta_\\theta(\\mathbf{x}_4) = \\gamma\n  (s_{21}\\cos\\theta-s_{22}\\sin\\theta,0,t_2)^\\top.\n\\end{align*}\n\nLet $\\underbar{x} < \\overline{\\mathrm{x}}$ be the $x$-coordinates of the end points of the base of\n$\\delta_\\theta(K)$.  The assumptions given in \\eqref{eq:tetra-param}\nyield\n\\begin{align*}\n   \\underbar{x} \\le 0, \\quad \\overline{\\mathrm{x}} = \\alpha\\cos\\theta, \\quad\n   0 \\le \\theta \\le \\frac{\\pi}{2}, \\qquad\n   \\underbar{x} =0, \\quad \\overline{\\mathrm{x}} \\ge \\alpha\\cos\\theta, \\quad\n   -\\frac{\\pi}{2} \\le \\theta \\le 0. \n\\end{align*}\nDefining\n$w = w(\\theta) := s_{21}\\cos\\theta - s_{22}\\sin\\theta$, $R_\\theta$ is\nwritten as\n\\[\n   R_\\theta = \\frac{1}{2\\gamma t_2}\n   \\left((\\overline{\\mathrm{x}} - \\gamma w)^2 + \\gamma^2t_2^2\\right)^{1/2}\n   \\left((\\underbar{x} - \\gamma w)^2 + \\gamma^2t_2^2\\right)^{1/2}.\n\\]\nTake an arbitrary\n$\\theta \\in \\left[-\\frac{\\pi}{3},\\frac{\\pi}{3}\\right]$.\nSuppose that the inequality\n\\begin{align}\n  \\gamma w \\le \\frac{\\underbar{x} + \\overline{\\mathrm{x}}}{2}\n  \\label{key-inequality1}\n\\end{align}\nholds and there exists a constant $C_1$ independent of\n$\\theta$ such that\n\\begin{align}\n |\\underbar{x} - \\gamma w| \\ge C_1 \\gamma \\mathbf{s}_2.  \\label{key-inequality2}\n\\end{align}\nWe then have\n\\begin{align*}\n   (\\underbar{x} - \\gamma w)^2 + \\gamma^2 t_2^2 & \\ge C_1^2\n   (1 - t_2^2) + \\gamma^2 t_2^2 \\ge\n   \\min\\{1, C_1^2\\} \\gamma^2, \\\\\n   (\\overline{\\mathrm{x}} - \\gamma w)^2 + \\gamma^2 t_2^2 & \\ge \\max\\left\\{\n      (\\underbar{x} - \\gamma w)^2 + \\gamma^2 t_2^2, \\;\n      \\frac{\\alpha^2\\cos^2\\theta}{4} + \\gamma^2 t_2^2 \\right\\} \\\\\n  & \\ge \\max\\left\\{\\min\\{1,C_1^2\\} \\gamma^2,\\, C_2^2\\alpha^2\\right\\} \\\\\n  & \\ge \\min\\{C_1^2,C_2^2\\}\\max\\{\\alpha^2,\\gamma^2\\},\n\\end{align*}\nwhere $C_2 = C_2(\\theta) := \\frac{\\cos\\theta}{2}$.\nHere, we have used the fact that $\\overline{\\mathrm{x}}/2 \\ge (\\overline{\\mathrm{x}}+\\underbar{x})/2 \\ge \\gamma w$ and\n$\\overline{\\mathrm{x}} - \\gamma w \\ge \\overline{\\mathrm{x}}/2 \\ge (\\alpha\\cos\\theta)/2$.  Hence,\nsetting\n\\[\n   \\frac{\\min\\{C_1,C_2\\}\\min\\{1,C_1\\}}{2\\sqrt{2}} \n  \\ge \\frac{\\min\\left\\{C_1,\\frac{1}{4}\\right\\}\n   \\min\\{1,C_1\\}}{2\\sqrt{2}} =: C_3,\n\\]\nwe obtain\n\\begin{align*}\n  R_P \\ge R_\\theta \\ge \\frac{\\sqrt{2}C_3}{t_2}\\max\\{\\alpha,\\gamma\\}\n   = \\frac{\\sqrt{2}C_3 \\max\\{\\alpha,\\gamma\\}}{\\sqrt{1-\\mathbf{s}_2^2}}\n   \\ge C_3 \\frac{\\max\\{\\alpha,\\gamma\\}}{\\sqrt{1-\\mathbf{s}_2}},\n\\end{align*}\nand the key inequality \\eqref{RP-est} is shown.\n\nFix $\\varphi$ such that\n\\[\n   0 < \\varphi < \\frac{\\pi}{6}, \\qquad \n   \\sin2\\varphi\\tan2\\varphi \\le \\frac{1}{6}.\n\\]\nIn the following, we will show that, according to \n$\\mathbf{x}_4 = (\\gamma s_{21}, \\gamma s_{22}, \\gamma t_2)^\\top$, we can\ntake an appropriate\n$\\theta \\in \\left[-\\frac{\\pi}{3},\\frac{\\pi}{3}\\right]$ such that\nconditions \\eqref{key-inequality1} and \\eqref{key-inequality2}\nhold with $C_1 = \\sin\\varphi$.  \n\n\\vspace{0.3cm}\n\\noindent\n\\textbf{Case 1.} Suppose that $|s_{22}|\\tan\\varphi \\le |s_{21}|$. \\\\\nIn this case, we set $\\theta = 0$ and have $\\underbar{x} = 0$, $\\overline{\\mathrm{x}} = \\alpha$,\nand $\\gamma w = \\gamma s_{21} \\le \\alpha/2 = (\\underbar{x} + \\overline{\\mathrm{x}})/2$\nbecause of \\eqref{eq:tetra-param}.  Hence, \\eqref{key-inequality1} holds.\nFor \\eqref{key-inequality2}, we note that\n\\begin{align*}\n   |\\underbar{x} - \\gamma w| = \\gamma|s_{21}| & = \\gamma\n   \\left(s_{21}^2\\sin^2\\varphi + s_{21}^2\\cos^2\\varphi\\right)^{1/2} \\\\\n  & \\ge \\gamma\n   \\left(s_{21}^2\\sin^2\\varphi + s_{22}^2\\sin^2\\varphi\\right)^{1/2}\n   = \\gamma \\mathbf{s}_2 \\sin\\varphi,\n\\end{align*}\nand so \\eqref{key-inequality2} holds with $C_1 := \\sin\\varphi$.\n\n\\vspace{0.3cm}\n\\noindent\n\\textbf{Case 2:} Suppose that $|s_{22}|\\tan\\varphi > |s_{21}|$\nand $3\\gamma s_{22} \\tan2\\varphi \\le \\alpha$. \\\\\nIn this case, set $\\theta = -2\\varphi$. We then have $\\underbar{x} = 0$\nand $\\overline{\\mathrm{x}} \\ge \\alpha\\cos\\theta = \\alpha \\cos2\\varphi$.  If\n$s_{22} > 0$, then\n\\begin{align*}\n \\gamma w & = \\gamma (s_{21}\\cos2\\varphi + s_{22}\\sin2\\varphi)\n  \\le \\gamma (s_{22}\\cos2\\varphi\\tan\\varphi + s_{22}\\sin2\\varphi) \\\\\n  & = \\frac{\\gamma s_{22}}{2}\n   (3 \\sin2\\varphi - 2\\sin^2\\varphi\\tan\\varphi) \\\\\n  & \\le \\frac{2\\gamma s_{22}}{2}\\sin2\\varphi\n   \\le \\frac{\\alpha \\cos2\\varphi}{2} \\le \\frac{\\underbar{x} + \\overline{\\mathrm{x}}}{2}.\n\\end{align*}\n   If $s_{22} \\le 0$, then\n\\begin{align*}\n \\gamma w & = \\gamma (s_{21}\\cos2\\varphi + s_{22}\\sin2\\varphi)\n  \\le \\gamma (- s_{22}\\cos2\\varphi\\tan\\varphi + s_{22}\\sin2\\varphi) \\\\\n  & = \\gamma s_{22}\\tan\\varphi \\le 0 \\le \\frac{\\underbar{x} + \\overline{\\mathrm{x}}}{2}.\n\\end{align*}\nThus, in either case, \\eqref{key-inequality1} holds.\nFor \\eqref{key-inequality2}, we note that\n\\begin{align*}\n   |\\underbar{x} - \\gamma w| & = \\gamma\n  |s_{21}\\cos2\\varphi + s_{22}\\sin2\\varphi| \\ge \\gamma \\left(\n  |s_{22}|\\sin2\\varphi - |s_{21}|\\cos2\\varphi\\right) \\\\\n  & \\ge \\gamma \\left(\n  |s_{22}|\\sin2\\varphi - |s_{22}|\\cos2\\varphi\\tan\\varphi\\right)\n  = \\gamma|s_{22}|\\tan\\varphi \\\\\n  & = \\gamma \\left(\n    s_{22}^2 \\sin^2\\varphi\\tan^2\\varphi + s_{22}^2\\sin^2\\varphi\n   \\right)^{1/2} \\\\\n  & = \\gamma \\left(s_{21}^2 \\sin^2\\varphi + s_{22}^2\\sin^2\\varphi\n   \\right)^{1/2} = \\gamma\\mathbf{s}_2\\sin\\varphi,\n  \\end{align*}\nand so \\eqref{key-inequality2} holds with $C_1 := \\sin\\varphi$.\n\n\\vspace{0.3cm}\n\\noindent\n\\textbf{Case 3:} Suppose that $|s_{22}|\\tan\\varphi > |s_{21}|$\nand $3\\gamma s_{22} \\tan2\\varphi > \\alpha$. \\\\\nIn this case, set $\\theta = 2\\varphi$.  We then have\n\\[\n   \\underbar{x} = \\min\\{\\eta\\cos2\\varphi - \\xi\\sin2\\varphi,0\\}, \\qquad\n    \\overline{\\mathrm{x}} = \\alpha \\cos2\\varphi.\n\\]\nIf $\\eta\\cos2\\varphi - \\xi\\sin2\\varphi \\le 0$, then\n\\begin{align*}\n   \\underbar{x} & = \\eta\\cos2\\varphi - \\xi\\sin2\\varphi\n   = \\alpha \\cos2\\varphi + (\\eta - \\alpha) \\cos2\\varphi\n      - \\xi\\sin2\\varphi \\\\\n  & \\ge \\alpha \\cos2\\varphi -\n   \\left((\\eta - \\alpha)^2 + \\xi^2\\right)^{1/2} \n   \\ge \\alpha \\cos2\\varphi - \\alpha = -2\\alpha \\sin^2\\varphi,\n\\end{align*}\nbecause the lengths of all edges of the base are less than $\\alpha$.\nEven if $\\eta\\cos2\\varphi - \\xi\\sin2\\varphi \\ge 0 = \\underbar{x}$, the above\ninequality obviously holds.  Because \n\\begin{align*}\n  \\gamma w & = \\gamma (s_{21}\\cos2\\varphi - s_{22}\\sin2\\varphi) \\\\\n  & \\le \\gamma (s_{22}\\cos2\\varphi\\tan\\varphi - s_{22}\\sin2\\varphi)\n  = -\\gamma s_{22}\\tan\\varphi,\n\\end{align*}\nwe have\n\\begin{align*}\n  \\underbar{x} - \\gamma w & \\ge -2\\alpha\\sin^2\\varphi + \\gamma s_{22}\\tan\\varphi \\\\\n  & \\ge -2\\alpha\\sin^2\\varphi + \\frac{\\alpha\\tan\\varphi}{6\\tan2\\varphi}\n + \\frac{\\gamma}{2} s_{22}\\tan\\varphi \\\\\n  & = \\alpha\\left(\\frac{1}{6}- \\sin2\\varphi\\tan2\\varphi\\right)\n   \\frac{\\tan\\varphi}{\\tan2\\varphi}\n   + \\frac{\\gamma}{2} s_{22}\\tan\\varphi\n    \\ge \\frac{\\gamma}{2} s_{22}\\tan\\varphi > 0.\n\\end{align*}\nTherefore,\n\\[\n   \\gamma w < \\underbar{x} < \\frac{\\underbar{x} + \\overline{\\mathrm{x}}}{2},\n\\]\nand \\eqref{key-inequality1} holds.\nFor \\eqref{key-inequality2}, we note that\n\\begin{align*}\n   |\\underbar{x} - \\gamma w| & \\ge \\frac{\\gamma}{2} s_{22}\\tan\\varphi \\\\\n  & = \\gamma \\left(\n    s_{22}^2 \\sin^2\\varphi\\tan^2\\varphi + s_{22}^2\\sin^2\\varphi\n   \\right)^{1/2} \\\\\n  & \\ge \\gamma \\left(s_{21}^2 \\sin^2\\varphi + s_{22}^2\\sin^2\\varphi\n   \\right)^{1/2} = \\gamma\\mathbf{s}_2\\sin\\varphi,\n  \\end{align*}\nand so \\eqref{key-inequality2} holds with $C_1 := \\sin\\varphi$.\n\n\\vspace{0.3cm}\nUsing inequalities \\eqref{RB-est} and \\eqref{RP-est}, we have shown the\nfollowing lemma.\n\n\\begin{lemma}\\label{lemma:geometric}\nLet $K$ be a tetrahedron with vertices given by \\eqref{eq:position}\nand the parameters in \\eqref{eq:tetra-param}.\nWe then have\n\\[\n  \\prod_{i=1}^2(1-\\mathbf{s}_i)^{-1/2} \\le C\n   \\frac{R_BR_P}{h_B\\max\\{\\alpha,\\gamma\\}},\n\\]\nwhere $C$ is a constant independent of $K$.\n\\end{lemma}\n\nTake a facet $B$ so that $R_BR_P/h_B = R_K$.\nCombining Theorem~\\ref{thmB} and Lemma~\\ref{lemma:geometric} with the \nprojected circumradius $R_K$, we have\n\\begin{align*}\n B_p^{m,k}(K) := \\sup_{v \\in \\mathcal{T}_p^k}\\frac{|v|_{m,p,K}}{|v|_{k+1,p,K}}\n  & \\le C  \\frac{\\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-m}}\n  {\\prod_{i=1}^2(1-\\mathbf{s}_i)^{m/2}} \\\\\n  & \\le C \\left(\\frac{R_BR_P}{h_B}\\right)^m\n    \\left(\\max\\{\\alpha,\\gamma\\}\\right)^{k+1-2m} \\\\\n  & \\le C R_K^m h_K^{k+1-2m},\n\\end{align*}\nwhere the constant $C = C(k,m,p)$ is independent of $K$\nwith vertices \\eqref{eq:position}.  Note that Sobolev (semi-)norms \nare affected by rotation up to a constant.  Therefore, we have\nderived the following result, which is the main theorem of this\npaper.\n\n\\begin{theorem}\\label{thm:main}\nLet $K$ be an arbitrary tetrahedron.  Let $h_K:=\\mathrm{diam}K$ and\n$R_K$ be the projected circumradius of $K$ defined by\n\\eqref{def:RK}.  Assume that $k$, $m$ are integers with $k \\ge 1$,\n$0 \\le m \\le k$, and $p$ $(1 \\le p \\le \\infty)$ is taken as in \\eqref{p-cond}.\nThen, for arbitrary $v \\in W^{k+1,p}(K)$,\nthere exists a constant $C = C(k,m,p)$ independent of $K$ such that\n\\begin{align}\n  |v - \\mathcal{I}_K^k v|_{m,p,K} \\le C R_K^m h_K^{k+1-2m}|v|_{k+1,p,K}\n   = C \\left(\\frac{R_K}{h_K}\\right)^m h_K^{k+1-m}|v|_{k+1,p,K}.\n  \\label{main-est}\n\\end{align}\n\\end{theorem}\nNote that the error estimation \\eqref{main-est} is exactly same as\n\\eqref{eq:circumradius-est}.\n\n\\section{Concluding remarks}\nIn this final section, we present some remarks on the newly obtained\nerror estimation.\n\n(1) For finite element error analysis, the most important\ncase is $p=2$ and $m=1$.  In this case, $k$ should be greater than or\nequal to $2$\nbecause of the restriction in \\eqref{p-cond}.  In this case, the main\nestimation is\n\\[\n   |v - \\mathcal{I}_K^k v|_{1,p,K} \\le C R_K h_K^{k-1}|v|_{k+1,p,K}\n   = C \\left(\\frac{R_K}{h_K}\\right) h_K^{k}|v|_{k+1,p,K}.\n\\]\n\nLet $\\Omega \\subset \\mathbb R^3$ be a bounded polygonal domain.  \nSuppose that we compute a numerical solution of the Poisson equation\n\\[\n    -\\Delta u = f \\text{ in } \\Omega, \\quad\n    u = 0 \\text{ on } \\partial\\Omega\n\\]\nby the finite element method with conforming simplicial elements.\nOne difficulty we may encounter in the numerical computation is the bad\ntriangulation of $\\Omega$ with many slivers \\cite{Sliver}.  Consider\nthe typical sliver $K$ mentioned in Section~\\ref{intro}, whose vertices are\n$(\\pm h,0,0)^\\top$ and $(0,\\pm h,h^\\alpha)^\\top$.  We can see that\n$R_K = \\mathcal{O}(h_K^{2-\\alpha})$ and\n$R_Kh_K^{k-1} = \\mathcal{O}(h_K^{k+1-\\alpha})$.  Thus, if we use\na $k$th-order conforming Lagrange finite element method with\ntriangulation that contains many slivers like $K$, the theoretical\nconvergence rate may be $\\mathcal{O}(h^{k+1-\\alpha})$.  This is\nworse than the expected rate $\\mathcal{O}(h^{k})$, but we can still\nexpect convergence if $k+1-\\alpha > 0$.   Therefore,\nbad triangulation with many slivers can be remedied by using\nhigher-order Lagrange elements.\n\n(2) Let $K$ be a tetrahedron with the inscribed sphere $S_K$.\nIf the regularity assumption is imposed on $K$, there exists\na constant $\\sigma$ such that $h_K/\\rho_K \\le \\sigma$, where\n$\\rho_K := \\mathrm{diam}S_K$.\nSuppose that $K$ is at the standard position.  Recall\n$\\delta_\\theta$, the projection introduced in\nSection~\\ref{section:Circumradius}.  It is clear that\n$\\delta_\\theta(S_K) \\subset \\delta_\\theta(K)$ and \n$\\mathrm{diam}\\delta_\\theta(S_K) = \\rho_K$.  Therefore, recalling that $R_P$\nis the maximum value of the circumradius of $\\delta_\\theta(K)$, there\nexists a constant $C$ such that $R_P \\le C h_K$.\nThe regularity assumption implies the maximum angle condition, and\nas was pointed out in \\cite[Section~4.1]{KobayashiTsuchiya3}, the\nmaximum angle condition of $B$ implies the boundedness of $R_B/h_B$.\nThus, Theorem~\\ref{thm:main} is an extension of the standard estimation\n\\eqref{eq:standard-est}.\n\n(3) For tetrahedrons, a (generalized) maximum angle condition was given\nby K\\v{r}\\'{i}\\v{z}ek \\cite{Krizek2}, \\cite{Duran}.\nA slightly more general condition has been stated by Jamet \\cite{Jamet}.\nThe authors conjecture that Theorem~\\ref{thm:main} is an extension\nof these results.\n\n(4) In Theorem~\\ref{thmB}, we obtained an error estimation for\nLagrange interpolation on tetrahedrons in terms of the singular values\n${\\prod_{i=1}^2(1-\\mathbf{s}_i)^{-1/2}}$ of the linear transformation.\nIn Theorem~\\ref{thm:main}, we showed that the projected circumradius\nis a geometric interpretation of the singular values.  The authors,\nhowever, are not completely sure whether the projected circumradius is\nthe \\textit{best} interpretation. \n\nFurther research on the geometry of tetrahedrons is required \nto ascertain the geometric properties of the singular values \n${\\prod_{i=1}^2(1-\\mathbf{s}_i)^{-1/2}}$ and the projected circumradius,\nand the relationship between these singular values and prior results.\n\n\n", "meta": {"timestamp": "2016-06-14T02:18:15", "yymm": "1606", "arxiv_id": "1606.03918", "language": "en", "url": "https://arxiv.org/abs/1606.03918", "abstract": "This paper describes the analysis of Lagrange interpolation errors on tetrahedrons. In many textbooks, the error analysis of Lagrange interpolation is conducted under geometric assumptions such as shape regularity or the (generalized) maximum angle condition. In this paper, we present a new estimation in which the error is bounded in terms of the diameter and projected circumradius of the tetrahedron. Because we do not impose any geometric restrictions on the tetrahedron itself, our error estimation may be applied to any tetrahedralizations of domains including very thin tetrahedrons.", "subjects": "Numerical Analysis (math.NA)", "title": "Error Analysis of Lagrange Interpolation on Tetrahedrons", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9848109534209825, "lm_q2_score": 0.8539127548105611, "lm_q1q2_score": 0.8409426342033263}}
{"url": "https://arxiv.org/abs/0901.3646", "title": "Fixed point and spectral characterization of finite dimensional C*-algebras", "abstract": "We show that the following conditions on a C*-algebra are equivalent: (i) it has the fixed point property for nonexpansive mappings, (ii) the spectrum of every self adjoint element is finite, (iii) it is finite dimensional. We prove that (i) implies (ii) using constructions given by Goebel, that (ii) implies (iii) using projection operator properties derived from the spectral and Gelfand-Naimark-Segal theorems, and observe that (iii) implies (i) by Brouwer's fixed point theorem.", "text": "\\section{Statement of Result}\n\\label{introduction}\n\\setcounter{equation}{0}\nWe let ${\\Bbb C},$ ${\\Bbb R},$ ${\\Bbb Q},$ ${\\Bbb Z},$ ${\\Bbb N} = \\{1,2,3,4,...\\},$ and\n${\\Bbb T} = \\{ \\, z \\in {\\Bbb C} \\, : \\, |z| = 1 \\, \\}$ denote the fields of\ncomplex, real, and rational numbers, the ring of integers, the set\nof natural numbers, and the circle group. If ${\\cal B }$ is a Banach\nspace and $K \\subset {\\cal B }$ then a map $M \\, : \\, K \\rightarrow K$\nis {\\it nonexpansive} if $||M(x) - M(y)|| \\leq ||x-y||, \\, x, y\n\\in K,$ and ${\\cal B }$ has the {\\it fixed point property for\nnonexpansive maps} (FPP) if $M$ has a fixed point whenever $M$ is\nnonexpansive and $K$ is a nonempty bounded closed convex subset of\n${\\cal B },$ \\cite{goebel}. Throughout this paper ${\\cal A }$ denotes a (not\nnecessarily unital) $C^*$--algebra with ${\\cal A } \\neq \\{ 0\\}$ and\n$\\widetilde {\\cal A }$ denotes ${\\cal A }$ if ${\\cal A }$ is unital and the\nunitization of ${\\cal A }$ otherwise and $I$ denotes its unit.\n$\\text{Inv}(\\widetilde {\\cal A })$ denotes the set of invertible\nelements in $\\widetilde {\\cal A }$ and the {\\it spectrum} of an element\n$A \\in {\\cal A }$ is defined by\n    $\\hbox{spec}(A) = \\{ \\, \\mu \\in {\\Bbb C} \\, : \\, \\mu \\, I - A \\notin\n    \\text{Inv}(\\widetilde {\\cal A })\\, \\},$\n\\cite{murphy}. We let ${\\cal A }_s, {\\cal A }_u,$ and ${\\cal A }_n$ denote the sets\nof self-adjoint, unitary, and normal elements in ${\\cal A }.$ The\nfollowing result relates fixed point and spectral properties of\n$C^*$--algebras.\n\n\\bigskip\n\\begin{theorem}\n\\label{pro:GNS} The following properties of a $C^*$--algebra ${\\cal A }$\nare equivalent:\n\\begin{enumerate}\n\\item ${\\cal A }$ has the FPP,\n\\item ${spec}\\, (A)$ is finite whenever $A \\in {\\cal A }_s,$\n\\item ${\\cal A }$ has finite dimension.\n\\end{enumerate}\n\\end{theorem}\n\\textbf{Proof.} Proposition \\ref{pro1} shows that property (1)\nimplies property (2), Proposition \\ref{pro2} shows that property\n(2) implies property (3), and Brouwer's Fixed Point Theorem\n(\\cite{goebel}, Theorem 7.1) shows that property (3) implies\nproperty (1).\n\\hfill$\\square$\\\\\n\\section{Derivation of Proposition \\ref{pro1}}\n\\label{derivations1} \\setcounter{equation}{0}\nThroughout this section $S$ denotes a compact subset of ${\\Bbb C}.$\nWe observe that set $C(S),$ consisting of continuous complex valued functions on $S$ and equipped with adjoint $A^* = \\overline A$ and norm $||A|| = \\max_{\\, s \\in S} |A(s)|,$ is a unital abelian $C^*$--algebra and that for every $p \\in S$ the subset $I_p(S),$ consisting of functions in $C(S)$ that vanish at $p,$ is an ideal of $C(S).$\nIf $A \\in {\\cal A }$ then $\\hbox{spec}(A)$ is a compact subset of the\ndisc of radius $||A||$ centered at $0.$ Furthermore, ${\\cal A }_s$ and\n${\\cal A }_u$ are subsets of ${\\cal A }_n,$ $A \\in {\\cal A }_s$ implies\n$\\hbox{spec}(A) \\subseteq [ \\, -||A||,||A|| \\, ],$ and $A \\in\n{\\cal A }_u$ implies $\\hbox{spec}\\, (A) \\subseteq {\\Bbb T}.$ For $A \\in {\\cal A },$\nthe $C^*$--subalgebra $C^*(A)$ of ${\\cal A }$ generated by $A$ and the\n$C^*$--subalgebra $C^*(A,I)$ of $\\widetilde {\\cal A }$ generated by $A$\nand $I$ are abelian iff $A \\in {\\cal A }_n.$\n\\begin{lem}\n\\label{abelian}\nIf $A \\in {\\cal A }_n$ and\n$z \\, : \\, \\hbox{spec} \\, (A) \\rightarrow {\\Bbb C}$\ndenotes the inclusion map then there exists an isometric $*$--isomorphism\n$\\phi \\, : \\, C^*(A,I) \\rightarrow C(\\hbox{spec} \\, (A))$\nsuch that $\\phi(A) = z$ and $\\phi(I) = 1.$ If $A \\in\n{\\cal A }_s$ then\n$0 \\notin \\hbox{spec} \\, (A) \\implies \\phi(C^*(A)) = C(\\hbox{spec} \\, (A))$\nand\n$0 \\in \\hbox{spec} \\, (A) \\implies \\phi(C^*(A)) = I_0(\\hbox{spec} \\, (A)).$\n\\end{lem}\n\\textbf{Proof.}\nThe Gelfand theorem for abelian $C^*$--algebras implies that the maximal ideal space $\\Omega(C^*(A,I))$ of $C^*(A,I),$ identified with the set of multiplicative linear functionals $\\omega \\, : \\, C^*(A,I) \\rightarrow {\\Bbb C}$ equipped with the weak$^*$ topology, is a compact Hausdorff space and that $C^*(A,I)$ is isomorphic to $C(\\Omega(C^*(A,I)))$ under the Gelfand transform $B \\rightarrow \\widehat B$ defined by $\\widehat {B}(\\omega) = \\omega(B).$ We observe that $\\widehat {A} \\, : \\, \\Omega(C^*(A,I)) \\rightarrow \\hbox{spec} \\, (A)$ is a homeomorphism and that the function $\\phi \\, : \\, C^*(A,I) \\rightarrow C(\\hbox{spec} \\, (A))$ defined by  $\\phi(B) = \\widehat {B} \\circ (\\widehat {A})^{-1}, \\, B \\in C^*(A,I),$ satisfies the asserted properties, see also (\\cite{murphy}, Theorem 2.1.13), (\\cite{strocchi}, Proposition 1.5.3). The second assertion follows from the Weierstrass approximation theorem.\n\\hfill$\\square$\\\\\n\n\\bigskip\nThe following result generalizes Goebel's Example 2.3 in\n\\cite{goebel}.\n\\begin{lem}\n\\label{goebel1}\nIf $S$ is infinite then $C(S)$ does not\nhave the FPP.\n\\end{lem}\n\\textbf{Proof.} Choose a limit point $p \\in S$ and define the set\n$$\n    K = \\{ \\, A \\in C(S) \\, : \\,\n    %\n    \\ 0 \\leq \\|A\\| \\leq 1, \\ A(p) = 1 \\, \\}.\n$$\nChoose $B \\in K$ with $|B(s)| < 1$ for $s \\neq p$ and define the\nmap\n$M_B \\, : \\, K \\rightarrow K$\nby\n$ M_B(x) =  BA.$\nThen $K$ is a bounded closed convex set and the map $M_B$ is\nnonexpansive and does not have a fixed point. \\hfill$\\square$\\\n\\begin{lem}\n\\label{goebel2}\nIf $S$ is infinite and $p \\in S$ then $I_p(S)$ does not have the FPP.\n\\end{lem}\n\\textbf{Proof.} If there exists a limit point $q \\in S$ such that\n$q \\neq p$ then an argument similar to that used in Lemma\n\\ref{goebel1} shows that $I_p(S)$ does not have the FPP. Otherwise\n$p$ is the only limit point of $S$ and then $I_p(S)$ is\nisomorphic, as a Banach space, to the Banach space $c_0$\nconsisting of all complex valued convergent sequences. Let $K$ be\nthe unit ball in $c_0.$ Then $K$ is a bounded close convex set and\nGoebel's Example 4.1 in \\cite{goebel} shows that the mapping\n$M \\, : \\, K \\rightarrow K$\ndefined by\n$M(A) = (1-||a||,a_1,a_2,...), \\ a \\in c_0$\nis nonexpansive and does not have a fixed point. Therefore $c_0,$\nand hence $I_p(S),$ does not have the FPP. \\hfill$\\square$\\\n\\begin{pro}\n\\label{pro1}\nIf ${\\cal A }$ has the FPP and $A \\in {\\cal A }$ then $\\hbox{spec}\\, (A)$ is\nfinite.\n\\end{pro}\n\\textbf{Proof.} Assume that $A \\in {\\cal A }_s$ and $\\hbox{spec}\\, (A)$\nis infinite. It suffices to show that ${\\cal A }$ does not have the FPP.\nLemma \\ref{abelian} shows that $C^*(A)$ is either isomorphic to\n$C(\\hbox{spec}\\, (A))$ or isomorphic to $I_0(\\hbox{spec}\\, (A)).$\nTherefore Lemma \\ref{goebel1} and Lemma \\ref{goebel2} show that\nthe subalgebra $C^*(A)$ of ${\\cal A },$ and hence ${\\cal A },$ does not have\nthe FPP. \\hfill$\\square$\\\n\\section{Projections}\n\\label{projections} \\setcounter{equation}{0}\nThe set of projections\n    ${\\cal P } = \\{ \\, P \\in {\\cal A }_s \\, : \\, P^2 = P \\, \\}$\nin ${\\cal A }$ admits the partial order\n    $P \\leq Q  \\mbox{ iff } PQ = QP = P.$\nWe write $P < Q$ if $P \\leq Q$ and $P \\neq Q.$ We say $P$ and $Q$ are orthogonal if $PQ = QP = 0$ and then write $P \\perp Q.$ If $P, Q \\in {\\cal P }$ then $P + Q \\in {\\cal P }$ iff $P \\perp Q.$ A projection $P$ is maximal if there does not exist a projection $Q$  such that $P < Q.$ If ${\\cal A }$ is unital then $I$ is the unique maximal projection. A projection $P \\in {\\cal P }$ is minimal if $P \\neq 0$ and there does not exist a nonzero $Q \\in {\\cal P }$ such that $Q < P.$ Otherwise $P-Q$ is a nonzero projection and $P = Q + (P-Q).$\n\\begin{lem}\n\\label{spec}\nIf $A \\in {\\cal A }_s,$ $A \\neq 0,$ and $\\hbox{spec}\\, (A)$ is finite,\nthen there exist pairwise orthogonal nonzero projections\n$P_1,...,P_d \\in {\\cal P } \\cap C^*(A)$\nand nonzero elements\n$\\lambda_1,...,\\lambda_d \\in \\hbox{spec}\\, (A)$\nsuch that\n    $A = \\sum_{j=1}^{d} \\lambda_j \\, P_j.$\nFurthermore, $\\overline {AH} = AH$ and there exists a $P_A \\in {\\cal P }$ such that $P_AH = AH.$\n\\end{lem}\n\\textbf{Proof.} By assumption $\\hbox{spec}\\, (A)$ is finite and\nsince $A \\neq 0$ it contains at least one nonzero element. Let $d\n\\in {\\Bbb N},$ let $\\lambda_1,...,\\lambda_d$ be the distinct nonzero\nelements in $\\hbox{spec}\\, (A),$ and define $\\pi_j \\in\nC(\\hbox{spec}\\, (A)), j = 1,...,d$ by $\\pi_j(\\lambda_i) = 1$ if\n$i=j$ and zero otherwise. Let $\\phi \\, : \\, C^*(A,I) \\rightarrow\nC(\\hbox{spec} \\, (A))$ be the isometric $*$--isomorphism in Lemma\n\\ref{abelian} and define $P_j = \\phi^{-1}(\\pi_j), j = 1,...,d.$\nThen Lemma \\ref{abelian} ensures that $P_1,..,P_d$ are pairwise\northogonal minimal projection operators in $C^*(A)$ and that their\nsum equals $A.$ The second assertion follows by defining $P_A =\n\\sum_{j=1}^{d} P_j.$ \\hfill$\\square$\\\n\\begin{rem}\n\\label{spectral}\nLemma \\ref{spec} is a special case of the spectral theorem for normal elements in a $C^*$--algebra \\cite{murphy} and for normal operators on a Hilbert space \\cite{halmos}. Furthermore, for each $j = 1,...,d,$\n    $P_j = AL_j(A)/\\lambda_j$ where $L_j$\nis the Lagrange interpolating polynomial of degree $d-1$ that satisfies\n    $L_j(\\lambda_i) = 1$ if $i=j$ and equals zero otherwise.\n\\end{rem}\n\\begin{lem}\n\\label{AN}\nIf $\\hbox{spec}\\,  \\, (A)$ is finite for every $A \\in {\\cal A }_s$ then\nevery sequence in ${\\cal P }$ that either strictly increases or\nstrictly decreases must terminate.\n\\end{lem}\n\\textbf{Proof.} Assume to the contrary that there exists an\ninfinite sequence $P_n, \\ n \\in {\\Bbb N}$ that is either strictly\nincreasing or strictly decreasing and construct the sequence of\nnonzero projections $Q_n \\in {\\cal P }$ by $Q_n = P_{n+1} - P_n$ if\n$P_n$ increases and $Q_n = P_{n} - P_{n+1}$ if $P_n$ decreases.\nThen $A_k = \\sum_{n=1}^{k} (1/n)Q_n, \\, k \\in {\\Bbb N}$ is a Cauchy\nsequence in ${\\cal A }_s$ that converge to $A \\in {\\cal A }_s$ and\n$\\hbox{spec}\\, (A) = \\{ \\, 1/n \\, : \\, n \\in {\\Bbb N} \\, \\} \\cup \\{ \\,\n0 \\, \\}$ is infinite. This contradiction concludes the proof.\n\\hfill$\\square$\\\n\\begin{cor}\n\\label{maxmin}\nIf $\\hbox{spec}\\, (A)$ is finite for every $A \\in {\\cal A }_s$ then\nfor every nonzero $P \\in {\\cal P }$ there exists a maximal $P_{max} \\in {\\cal P }$ and\na minimal $P_{1} \\in {\\cal P }$ such that $P_{1} \\leq P \\leq P_{max}.$ Furthermore, there exist $d \\in {\\Bbb N}$ and nonzero pairwise orthogonal\nminimal $P_1,...,P_d \\in {\\cal P }$ such that $P = \\sum_{j=1}^{d} P_j.$\n\\end{cor}\n\\textbf{Proof.} The first assertion follows since otherwise there\nexists an infinite sequence of projections that either strictly\nincrease or strictly decreases thus contradicting Lemma \\ref{AN}.\nThe second assertion follows by recursively applying the first\nassertion to the projection $P-P_1$ whenever $P-P_1 \\neq 0.$\n\\hfill$\\square$\\\nFor any Hilbert space $H$ the set ${\\cal B }(H),$ consisting of bounded\noperators on $H,$ forms a $C^*$--algebra under the adjoint and\noperator norm. A $C^*$--subalgebra ${\\cal A }$ of ${\\cal B }(H)$ is called nondegenerate\nif $\\overline {{\\cal A } \\, H} = H.$\n\\begin{lem}\n\\label{GNS}\nEvery $C^*$ algebra ${\\cal A }$ is isomorphic to a nondegenerate $C^*$--subalgebra\n${\\cal A }$ of ${\\cal B }(H)$ for some Hilbert space $H.$\n\\end{lem}\n\\textbf{Proof.} This isomorphism is given by the\nGelfand-Naimark-Segal construction \\cite{murphy}.\n\\hfill$\\square$\\\\\n\n\\medskip\nThrougout the remainder of this paper we identify ${\\cal A }$ with a\nnondegenerate $C^*$--subalgebra of ${\\cal B }(H)$ for some Hilbert space $H$ as justified by Lemma \\ref{GNS}. If $H_1$ and $H_2$ are closed subspaces of $H$ that are orthogonal we write $H_1 \\perp H_2$ and if in addition their sum equals $H$ we write $H = H_1 \\bigoplus H_2$ and $H_2 = H_{1}^{\\perp}.$ We observe that for every $P, Q \\in {\\cal P },$ $PH$ is a closed subspace of $H,$ $P$ is the operator that orthogonally projects vectors onto $PH,$ $\\hbox{ker}\\, (P) = (PH)^{\\perp},$ and $P \\perp Q$ iff $PH \\perp QH.$ We also observe that for every $A \\in {\\cal A }_s,$\n    $H = \\hbox{ker}\\, (A) \\bigoplus \\overline {AH}.$\n\\begin{cor}\n\\label{unital}\nIf $\\hbox{spec}\\, \\, (A)$ is finite for every $A \\in {\\cal A }_s$ then ${\\cal A }$ is unital.\n\\end{cor}\n\\textbf{Proof.} Lemma \\ref{spec} and Corollary \\ref{maxmin} ensure\nthat there exists a maximal projection $P \\in {\\cal P }.$ It suffices to\nshow that $PH = H.$ Assume to the contrary that $PH \\neq H.$ Since\n${\\cal A }$ is nondegenerate there exists $A \\in {\\cal A }$ such that $AH$ is\nnot a subset of $PH$ and hence\n$\\hbox{ker} (P)$ is not a subset of $\\hbox{ker} (A) = \\hbox{ker} (A^*A).$\nTherefore, since $Q \\in {\\cal A }_s$ defined by\n    $Q = P + A^*A$\nsatisfies\n    $\\hbox{ker} (Q) = \\hbox{ker} (P) \\bigcap \\hbox{ker} (A),$\nit follows that $\\hbox{ker} (Q)$ is a proper subset of $\\hbox{ker} (P)$\nand hence\n    $PH$\nis a proper subspace of\n    $\\overline {QH}.$\nLemma \\ref{spec} implies that $\\overline {QH} = QH$ and that there\nexists $T \\in {\\cal P }$ such that $TH = QH.$ This implies that $P < T$\nthus contradicting the assumption that $P$ is maximal.\n\\hfill$\\square$\\\n\\section{Derivation of Proposition \\ref{pro2}}\n\\label{derivations2} \\setcounter{equation}{0}\nIf $H_1$ and $H_2$ are closed subspaces of $H$ we let $Hom_{{\\cal A }}(H_1,H_2)$ denote the subspace consisting of $A \\in {\\cal A }$ such that $AH_1 \\subset H_2$ and such that the restriction of $A$ to $H_{1}^{\\perp}$ equals $0.$ An element $U \\in {\\cal A }$ is called a partial isometry if $U^*U \\in {\\cal P }.$\n\\begin{lem}\n\\label{polar}\nEvery $A \\in {\\cal A }$ can be factored uniquely as\n$A = U|A|$\nwhere\n$H_1 = \\left(\\hbox{ker} (A)\\right)^{\\perp},$\n$H_2 = \\overline {AH},$\n$|A| \\in Hom_{{\\cal A }}(H_1,H_1)$ is the semipositive operator such that $|A|^2 = A^*A,$\nand\n$U \\in Hom_{{\\cal A }}(H_1,H_2)$ is a surjective partial isometry such that $U^*U$ is orthogonal projection onto $H_1.$\n\\end{lem}\n\\textbf{Proof.} This is the polar decomposition theorem\n\\cite{murphy}.\\hfill$\\square$\\\\\n\n\\medskip Throughout this remainder of this section we will assume\nthat ${\\cal A }$ satisfies the property that $A \\in {\\cal A }_s$ implies that\n${spec}\\, (A)$ is finite. Therefore Lemma \\ref{spec} and Lemma\n\\ref{polar} ensures that $AV$ is closed whenever $V$ is a closed\nsubspace of $H$ and $A \\in {\\cal A }.$\n\\begin{lem}\n\\label{dim1}\n    If $P \\in {\\cal P }$ is minimal and $A \\in Hom_{{\\cal A }}(PH, PH)$ then there exists\n    $c \\in {\\Bbb C}$ such that $A = cP.$ Therefore, since $P$ is a nonzero element\n    in $Hom_{{\\cal A }}(PH, PH),$ it follows that\n$\\dim \\left( Hom_{{\\cal A }}(PH, PH) \\right) = 1.$\n\\end{lem}\n\\textbf{Proof.}\n    Since $A_r = \\frac{1}{2}\\left(A + A^*\\right)$ and $A_i = \\frac{1}{2i}\\left(A - A^*\\right)$ are in ${\\cal A }_s \\cap Hom_{{\\cal A }}(PH, PH),$ Lemma \\ref{spec} implies that $A_r$ and $A_i$ are either zero or linear combinations having nonzero real coefficients of projections. Each projection is in $Hom_{{\\cal A }}(PH, PH)$ and therefore is $\\leq P.$ Since $P$ is minimal each projection must vanish or equal $P.$ Since $A = A_r + iA_i$ the proof is complete.\n\\hfill$\\square$\\\n\\begin{cor}\n\\label{dim2}\n    If $P, Q \\in {\\cal P }$ are minimal then\n    $\\dim \\left( Hom_{{\\cal A }}(PH, QH) \\right) \\in \\{0,1\\}.$\n\\end{cor}\n\\textbf{Proof.}\n    If nonzero elements $A, B \\in Hom_{{\\cal A }}(PH,QH)$ have polar decompositions $A = U|A|$ and $B = O|B|$ then $O^*U \\in Hom_{{\\cal A }}(PH,QH)$ and Lemma \\ref{dim1} implies that $|A|,$ $|B|,$ and $O^*U$ are multiples of $P.$ This completes the proof.\n\\hfill$\\square$\\\n\\begin{pro}\n\\label{pro2}\n${\\cal A }$ is finite dimensional.\n\\end{pro}\n\\textbf{Proof.} Corollary \\ref{unital} and Lemma \\ref{maxmin}\nimply that $I \\in {\\cal A }$ and that there exists minimal projections\n$P_1,...,P_d$ such that $I = \\sum_{j=1}^{d} P_j.$ Therefore every\n$A \\in {\\cal A }$ can be expressed as the sum\n$A = \\sum_{i,j=1}^{d} A_{ij}$\nwhere\n$A_{ij} = P_i A P_j \\in Hom_{{\\cal A }}(P_jH,P_iH).$\nTherefore\n${\\cal A } = \\bigoplus_{i,j = 1}^{d} Hom_{{\\cal A }}(P_jH,P_iH)$\nand hence Corollary \\ref{dim2} implies that $\\dim {\\cal A } \\leq d^2.$\n\\hfill$\\square$\\\n\\begin{rem}\n\\label{matrep}\nIn Corollary \\ref{dim2} we can define an equivalence relation on the set $\\{1,...,d\\}$ by $i \\sim j$ iff $\\dim \\left( Hom_{{\\cal A }}(P_jH,P_iH) \\right) = 1.$ This partitions the set into $n \\in {\\Bbb N}$ equivalence classes having cardinalities $d_1,...,d_n \\in {\\Bbb N}$ with $d = \\sum_{k = 1}^{n} d_k.$ Therefore ${\\cal A }$ is isomorphic to\n$\\bigoplus_{k = 1}^{n} {\\cal B }({\\Bbb C}^d),$\nwhich is a sum of matrix algebras, and\n${\\cal A }$ is abelian iff each $d_k = 1.$\n\\end{rem}\n\\section{Conclusions and Future Research}\n\\label{conclusions} \\setcounter{equation}{0}\nThis paper derives simple criteria for a $C^*$--algebra to be\nfinite dimensional and describes the structure of the algebra\nusing properties of their projections that are related to fixed\npoint and spectral theory. A sharpened understanding of these\nproperties may be useful in quantum mechanics where\n$C^*$--algebras provide the core mathematical foundation\n\\cite{strocchi}. Future research will examine the geometry of more complicated topological algebras that include the approximately finite dimensional $C^*$--algebras defined by Bratteli \\cite{bratteli} and classified by Elliott \\cite{elliott}, and the abelian Banach $*$--algebras whose fixed point properties were studied by\nDhompongsa and Fupinwong \\cite{dhompongsa-fupinwong}.\n\\\\ \\\\\n{\\it Acknowledgments} The authors thank Kazimierz Goebel for\nhelpful discussions.\n", "meta": {"timestamp": "2009-01-23T11:46:06", "yymm": "0901", "arxiv_id": "0901.3646", "language": "en", "url": "https://arxiv.org/abs/0901.3646", "abstract": "We show that the following conditions on a C*-algebra are equivalent: (i) it has the fixed point property for nonexpansive mappings, (ii) the spectrum of every self adjoint element is finite, (iii) it is finite dimensional. We prove that (i) implies (ii) using constructions given by Goebel, that (ii) implies (iii) using projection operator properties derived from the spectral and Gelfand-Naimark-Segal theorems, and observe that (iii) implies (i) by Brouwer's fixed point theorem.", "subjects": "Operator Algebras (math.OA); Functional Analysis (math.FA)", "title": "Fixed point and spectral characterization of finite dimensional C*-algebras", "lm_name": "Qwen/Qwen-72B", "lm_label": "1. YES\n2. YES", "lm_q1_score": 0.9907319855244919, "lm_q2_score": 0.8479677622198946, "lm_q1q2_score": 0.8401087847248764}}