chunk_uid
stringlengths
40
40
chunk_type
stringclasses
2 values
chunk_index
int64
0
6.71k
total_chunks
int64
1
6.71k
section_title
stringlengths
1
157
embed_text
stringlengths
1
83.3k
spans
dict
paper_doi
stringlengths
0
63
paper_id_arxiv
stringlengths
9
16
title
stringlengths
7
245
authors
listlengths
1
768
categories
listlengths
1
7
year
int64
2k
2.02k
language
stringclasses
2 values
discipline
stringclasses
8 values
dense_vector
listlengths
1.02k
1.02k
58beb316c7ffe6a68746e2ebd3689ecffd94dcf9
subsection
21
77
Polynomials of degree
Then, by Lemma REF , we haved(k(m)+1,m) > 1 + c_4(k(m)+1,m).If we let s=q_1q_2\cdots q_{k(m)} and t=q_{k(m)+1}\cdots q_m we have 2\phi (t)/t=d(k(m)+1,m), andc_4(k(m)+1,m)&=&8\cdot \sqrt{\frac{q_1q_2\cdots q_{k(m)}}{q_{k(m)+1}q_{k(m)+2}\cdots q_m}}\\ &=&\frac{8s}{\sqrt{q_1q_2\cdots q_m}}\ge \frac{8s}{\sqrt{p-1}}Since s ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.2140/pjm.1982.98.123", "end": 427, "openalex_id": "https://openalex.org/W1966377157", "raw": "D. J. Madden and W. Y. Vélez, Polynomials that represent quadratic residues at primitive roots, Pacific J. Math. 98 (1982), 123–137.", "s...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.009422050788998604, -0.015319415368139744, -0.05712738260626793, -0.030318403616547585, -0.014381024986505508, -0.01404534000903368, 0.019347628578543663, 0.008773569017648697, 0.024260826408863068, 0.01049776654690504, -0.030196337029337883, 0.0007958199712447822, -0.04186899960041046, ...
291e1b3ee39d2285f15bd27e96f8e89c111ea687
subsection
22
77
Polynomials of degree
Moreover, amongst these exceptions only for (p,k) \in \lbrace (13,1),(37,28),(61,18)\rbrace there exists \bar{\xi }\in S^*\cap S^*+1 such that k=- 2(1-2\bar{\xi }). In particular, \bar{\xi }=4 for (p,k)=(13,1), \bar{\xi }=26 for (p,k)=(37,28), and \bar{\xi }=5 for (p,k)=(61,18).
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.012751400470733643, -0.0005007843719795346, -0.04847211018204689, -0.0073410216718912125, -0.03317958489060402, -0.02428184077143669, 0.026632189750671387, 0.026647450402379036, -0.005437087267637253, -0.003164957044646144, -0.016177719458937645, -0.006020858883857727, -0.0239155534654855...
44cfbc50c8a14707c6f157f5b45f95b4453043e5
subsection
23
77
Vertex-transitive graphs of order
The goal of this paper is to prove that the Petersen graph is the only connected vertex-transitive graph of order a product of two primes without a Hamilton cycle. Recall that vertex-transitive graphs of prime-squared order are necessarily Cayley graphs of abelian groups. The existence of Hamilton cycles in such graphs...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 361, "openalex_id": "", "raw": "D. Marušič, Hamilonian circuits in Cayley graphs, Discrete Math. 46 (1983), 49–54.", "source_ref_id": "edb938b56f5864faa84d2c57544f61cc150fdfc3", "start": 273 }, { "arxiv_i...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.03359892964363098, -0.027068346738815308, -0.016418004408478737, 0.0037421148736029863, 0.026076551526784897, 0.01047486998140812, 0.01847788505256176, 0.02102603204548359, 0.007537634111940861, 0.016524814069271088, -0.014030070044100285, -0.014106362126767635, 0.00103566178586334, -0....
d0aa9f21ba075d10f2b218ef587d5062b893bfbd
subsection
24
77
Vertex-transitive graphs of order
If \gcd (c,q)=1, where c=a/\gcd (a,q) and a is the order of \alpha \in \mathbb {Z}_p^*, then it follows by that X is a Cayley graph of the group \langle \rho ,\sigma ^c\rangle =\langle \rho \rangle \rtimes \langle \sigma ^c\rangle . Thus in this case the result about existence of Hamilton cycles in Cayley graphs of sem...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.4153/cjm-1982-020-8", "end": 233, "openalex_id": "https://openalex.org/W2333628352", "raw": "B. Alspach and T. D. Parsons, A construction for vertex-transitive graphs, Canad. J. Math. 34 (1982), 307–318.", "source_ref_id": "2b23bb6...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.019212845712900162, 0.0037597815971821547, -0.040226418524980545, 0.01605394296348095, 0.03949391841888428, -0.039799127727746964, 0.026171589270234108, 0.030093511566519737, 0.012711914256215096, 0.006298730615526438, -0.009827698580920696, -0.02775867097079754, -0.003975334577262402, ...
beb29c56fc38620b2800324cf1237c64b7ee22d7
subsection
25
77
Vertex-transitive graphs of order
Since the above actions are primitive and hence the corresponding basic orbital graphs are connected, it suffices to prove the existence of Hamilton cycles solely in basic orbital graphs of these actions. This is done in the next three sections.Graphs arising from the actions in the first four rows and the last row of ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1016/0012-365x(89)90157-x", "end": 777, "openalex_id": "https://openalex.org/W2063575690", "raw": "B. Alspach, Lifting Hamilton cycles of quotient graphs, Discrete Math. 78 (1989), 25–36.", "source_ref_id": "7c0067fe920d8045ec54c42...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.017318572849035263, -0.013015632517635822, -0.023956088349223137, 0.04058802127838135, 0.00957480538636446, -0.017867883667349815, 0.02769445814192295, 0.05252029001712799, 0.007942129857838154, -0.0025329366326332092, -0.036712322384119034, -0.0009822759311646223, -0.026901008561253548, ...
44e2a9396f2383f02bfb8922848764128c6caaa5
subsection
26
77
Vertex-transitive graphs of order
Since the quotient graph X_\rho has a Hamilton cycle containing a double edge and since 13 is a prime, this cycle lifts to a Hamilton cycle in the original graph X (see Figure REF ). [Figure: A vertex-transitive graph arising from theaction of \textrm {PSL}(2,13) on cosets of D_{12}given in Frucht's notation with resp...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.043703097850084305, 0.012474609538912773, -0.0328688807785511, 0.03173967823386192, 0.007419913075864315, -0.045869939029216766, 0.017395785078406334, 0.054628871381282806, 0.02201940305531025, 0.003660363145172596, -0.04498489201068878, -0.030045880004763603, 0.0017634214600548148, -0....
4239f2892d7c6b522cbf126193b8305fc6d7f753
subsection
27
77
Vertex-transitive graphs of order
Let \mathcal {P}= \lbrace S_1, S_2, \ldots , S_m\rbrace be the set of orbits of \rho , and let \pi : X \rightarrow X_\mathcal {P} be the corresponding projection of X to its quotient X_\mathcal {P}. For a (possibly closed) path W = S_{i_1}S_{i_2}\ldots S_{i_k} in X_\mathcal {P} we let the lift of W be the set of all p...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1016/0012-365x(82)90220-5", "end": 429, "openalex_id": "https://openalex.org/W2917646739", "raw": "D. Marušič and T. D. Parsons, Hamiltonian paths in vertex-symmetric graphs of order 5p, Discrete Math. 42 (1982), 227–242.", "source...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.036715906113386154, -0.0002598993305582553, -0.025270799174904823, 0.02751404047012329, -0.0232259389013052, -0.015214364044368267, 0.03216838464140892, 0.019166741520166397, 0.01849529519677162, 0.017289742827415466, -0.004867985844612122, 0.0022241659462451935, 0.012749850749969482, -...
99ee6007497ef90983d5ad3e1f698c5888a17f94
subsection
28
77
Graphs arising from certain small rank/degree group actions
We deal here with the first four rows and the last row of Table REF of which the first three rows correspond to groups of rank 3 and 4.In the first proposition we show the existence of Hamilton cycles in the graphs arising from the first three rows of Table REF . With the exception of one of the graphs arising from the...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 2696, "openalex_id": "", "raw": "A. T. Balaban, Chemical Graphs, Part XIII; Combinatorial Patterns. Rev. Roumaine Math. Pures Appl. 17 (1972), 3–16.", "source_ref_id": "69ab62e9bfdbc10f808f22805460d2a2b6c71bc4", "start...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.0004268609336577356, -0.0005656503490172327, -0.017444541677832603, 0.019489658996462822, 0.03540799394249916, -0.029913650825619698, 0.028494277969002724, 0.0439547523856163, 0.019001273438334465, -0.006829774938523769, -0.028356919065117836, 0.005940759554505348, 0.008523864671587944, ...
eeb8de59c61222f1a795380a253345ca5cafec77
subsection
29
77
Graphs arising from certain small rank/degree group actions
More precisely, one suborbit is of length 6, one of length 10, two of length 12, four of length 20, five of length 30 and 27 of length 60. Of the latter 8 are non-self-paired. In Figure REF the graph of valency 6 is shown together with a Hamilton cycle in the quotient graph with respect to a semiregular automorphism o...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1006/jsco.1996.0125", "end": 579, "openalex_id": "https://openalex.org/W1976677460", "raw": "W. Bosma, J. Cannon, and C. Playoust, The Magma Algebra System I: The User Language, J. Symbolic Comput. 24 (1997), 235–265.", "source_ref...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.001579881296493113, -0.023096585646271706, -0.03813835233449936, 0.010632973164319992, 0.0241186935454607, -0.013157731853425503, 0.010701621882617474, 0.02655954845249653, -0.00042619608575478196, 0.00246373750269413, -0.03612464666366577, -0.004324889276176691, 0.011616942472755909, 0....
c8c6dc2ca8a9af2aa24a2ff79fc98df9f9feddee
subsection
30
77
Actions of
In this section the existence of Hamilton cycles in basic orbital graphs arising from the group action \hbox{\rm PSL}(2,q^2) on the cosets of \hbox{\rm PGL}(2,q) given in Row 5 of Table REF is considered. The following group-theoretic result due to Manning will be needed in this respect.Proposition 6.1 Let G be a tr...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 394, "openalex_id": "https://openalex.org/W2798943694", "raw": "H. Wielandt, Permutation groups, Academic Press, New York, 1966.", "source_ref_id": "8a2cf968027c31fecc14e57ca6fb71bcd0903446", "start": 289 } ] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.052935175597667694, -0.018962081521749496, -0.029976259917020798, 0.01778743974864483, 0.028069373220205307, -0.03209671750664711, 0.018321366980671883, 0.006296540144830942, -0.03407987952232361, 0.015972083434462547, -0.023538609966635704, -0.02649809792637825, 0.01999942772090435, 0....
560772a7b6ddd7f6247693117ace5100c10110e1
subsection
31
77
Actions of
(Note that there is only one conjugacy class of subgroups in G isomorphic to N.) A subgroup of N is isomorphic to one of the following groups \mathbb {Z}_q^2\rtimes \mathbb {Z}_{q-1}, \mathbb {Z}_q, \mathbb {Z}_q\rtimes \mathbb {Z}_{q-1}, \mathbb {Z}_q\rtimes \mathbb {Z}_l, where 2\le l<q-1, and \mathbb {Z}_l, where l ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.023662537336349487, 0.025813676416873932, -0.008307061158120632, -0.0034174134489148855, 0.006018614396452904, -0.02491355501115322, 0.013997664675116539, -0.015614833682775497, -0.03246542811393738, 0.015004580840468407, -0.02029089257121086, -0.031183898448944092, 0.0027957188431173563,...
8a6c1d3663c62363f3b12eed532338d0d6e6d769
subsection
32
77
Actions of
Since |\mathcal {H}|=q(q^2+1)/2 and since the length of an orbit of N on \mathcal {H} with coset stabilizer isomorphic to \mathbb {Z}_2 or to \mathbb {Z}_1 equals, respectively, \frac{q^2(q-1)}{2} and q^2(q-1), we have\frac{q(q^2+1)}{2} = q\frac{q+1}{2} + a\frac{q^2(q-1)}{2} + bq^2(q-1),where a is the number of orbits ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.061632752418518066, 0.01298897247761488, -0.03208322077989578, 0.013256078585982323, 0.026695314794778824, -0.03168637678027153, -0.012302129529416561, 0.007997910492122173, -0.0033579012379050255, 0.006028959061950445, -0.02110898867249489, 0.00815817341208458, 0.000002504116764612263, ...
30f867f9079f28eb6a3deed179be664f5e9020c2
subsection
33
77
Actions of
Formal proofs are given in Propositions REF and REF .It follows from the previous paragraph that we only need to prove that the subgraph of X_\mathcal {P} induced on the large orbit of N contains a Hamilton cycle with at least one double edge in the corresponding multigraph or two components each of which contains a Ha...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 945, "openalex_id": "", "raw": "C. E. Praeger and M. Y. Xu, Vertex primitive transitive graphs of order a product of two distinct primes, J. Combin. Theory Ser. B 59 (1993), 245–266.", "source_ref_id": "5dfe58ee8c60a5542d538...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.057547200471162796, -0.011434683576226234, -0.0218166746199131, 0.010862568393349648, 0.0006236059707589447, -0.037622321397066116, 0.031245138496160507, -0.0009802246931940317, -0.01694987714290619, 0.0038560591638088226, -0.035883091390132904, -0.027583599090576172, 0.01649218611419201,...
99bacdfbc15ae1f4cf048c21994151f7a439965b
subsection
34
77
Actions of
One can now see that \cdot \colon W\otimes \overline{W}\times \hbox{\rm SL}(2,q^2) \rightarrow W\otimes \overline{W} defined by the rule(\mathbf {w}\otimes \mathbf {w}^{\prime })\cdot g= \mathbf {w}g\otimes \mathbf {w}^{\prime } * g= \mathbf {w}g\otimes \mathbf {w}^{\prime }g^\phiis an action of \hbox{\rm SL}(2,q^2) on...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.02089472860097885, 0.015621442347764969, 0.011164714582264423, -0.032387591898441315, 0.022543108090758324, -0.017491132020950317, 0.01877320371568203, 0.01645326241850853, -0.006509570404887199, 0.026267223060131073, 0.01087472215294838, -0.0058265612460672855, 0.02342834696173668, -0....
3ceccf63eade17982cd53ca360ff77f53f5577fb
subsection
35
77
Actions of
It follows that G fixes a non-degenerate symmetric bilinear form of W\otimes \overline{W} defined by the rule(\mathbf {w}_1^{\prime }\otimes \mathbf {w}_2^{\prime }, \mathbf {w}_1^{\prime \prime }\otimes \mathbf {w}_2^{\prime \prime })= f(\mathbf {w}_1^{\prime }, \mathbf {w}_1^{\prime \prime })f(\mathbf {w}_2^{\prime }...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.014080964960157871, 0.017864366993308067, -0.021113820374011993, -0.041739460080862045, 0.008535536006093025, 0.008032099343836308, 0.03374549746513367, -0.001913630636408925, 0.021159587427973747, 0.006720114033669233, -0.013455483131110668, -0.020106947049498558, 0.01623201183974743, ...
0a147356f81bd706a7a3f0da8cac2c87275bd840
subsection
36
77
Actions of
The following result characterizing suborbits of the action of G on the cosets of \hbox{\rm PGL}(2,q) in the context of the action of P\Omega ^{-}(4, q) on \Omega was proved in .Proposition 6.3 For any \langle \mathbf {v}\rangle \in \Omega , the nontrivial suborbits of the action of G on \Omega (that is, the orbits ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 179, "openalex_id": "", "raw": "C. E. Praeger and M. Y. Xu, Vertex primitive transitive graphs of order a product of two distinct primes, J. Combin. Theory Ser. B 59 (1993), 245–266.", "source_ref_id": "5dfe58ee8c60a5542d538...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.00785884540528059, 0.024293752387166023, -0.026964232325553894, 0.0007105388212949038, 0.014184071682393551, -0.04602384194731712, 0.018464472144842148, -0.0012732474133372307, 0.005878874100744724, -0.008972818031907082, -0.010468287393450737, -0.010109679773449898, 0.022294705733656883, ...
4d1000f341ad171cd2bc8dc1ceb6343d5ce28c34
subsection
37
77
Actions of
For k\in F_q we have\mathbf {v}_1\rho ^k&=&\mathbf {v}_1,\\ \mathbf {v}_2\rho ^k&=&k^2\mathbf {v}_1+\mathbf {v}_2+k\mathbf {v}_3,\\ \mathbf {v}_3\rho ^k&=&2k\mathbf {v}_1+\mathbf {v}_3,\\ \mathbf {v}_4\rho ^k&=&\mathbf {v}_4,and so \rho ^k maps the vector x=\sum _{i=1}^4x_i\mathbf {v}_i\in V to\mathbf {x}\rho ^k=(x_1+k...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.019775526598095894, 0.023361368104815483, -0.04180937260389328, 0.0007576994830742478, 0.0006938028382137418, 0.015823474153876305, 0.03869655728340149, 0.009544438682496548, -0.005031620617955923, 0.03108236752450466, -0.040161412209272385, 0.013534639030694962, 0.01197823230177164, 0....
fa7072b908ed077e255baa857d87e4c12ee66db8
subsection
38
77
Actions of
By \mathcal {I}_P and \mathcal {L}_P, we denote, respectively, the set of blocks in \mathcal {I} and \mathcal {L}; that is, \mathcal {P}=\mathcal {I}_P\cup \mathcal {L}_P.Remark 6.4 Note thatN=N_G(P)=\langle \left[ \begin{array}{cc} a&b\\0&a^{-1} \end{array}\right]\bigm |a\in \langle \alpha \rangle , b\in F_{q^2}\rangl...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.057599987834692, 0.03437073901295662, -0.028647370636463165, 0.02692272886633873, -0.004769474267959595, -0.023946575820446014, -0.00002116454197675921, 0.011179647408425808, 0.005490618757903576, 0.012690617702901363, -0.0251522995531559, -0.020161522552371025, 0.03476756066083908, 0.0...
61dd1a9237fb18537b33e287d908442cd3c4d941
subsection
39
77
Actions of
Then y_2\ne 0 and x_3\ne 0, and \langle \mathbf {x}\rangle \sim \langle \mathbf {y}\rho ^k\rangle if and only if(\mathbf {x},\mathbf {y}\rho ^k)=((0,0,x_3,x_4), (y_1+k^2y_2,y_2,ky_2, y_4))=\pm 2\lambda ,that is, if and only if-2x_3ky_2+2\theta x_4y_4=\pm 2\lambda .From (REF ) we get that k=\frac{\theta x_4y_4\mp \lambd...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 1152, "openalex_id": "", "raw": "T. Tchuda, Combinatorial-geometric characterization of some primitive representations of the groups \\hbox{\\rm PSL}_n(q), n=2,3, PhD Thesis, Kiev, 1986 (in Russian).", "source_ref_id": "86ba...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.029068170115351677, 0.014373866841197014, -0.01425942499190569, -0.01362618152052164, 0.041839852929115295, -0.05426058545708656, 0.025665439665317535, 0.026306312531232834, 0.01158149167895317, 0.02734391577541828, -0.022842546924948692, -0.018081774935126305, -0.0025940099731087685, -...
b625a3ebea4f01c363ed3e212a93431335b3415b
subsection
40
77
Actions of
We may choose H to consist of all the matrices of the form\left[ \begin{array}{cc} x & 0 \\ 0 & x^{-1} \end{array} \right] \ (x\in F^*) \ \textrm { and } \ \left[ \begin{array}{cc} 0 & -x \\ x^{-1} & 0 \end{array} \right] \ (x\in F^*).Note that, since p\ge 13, H is a dihedral subgroup D_{p-1}. Further, letg= \left[ ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 1333, "openalex_id": "", "raw": "D. Marušič and R. Scapellato, A class of non-Cayley vertex-transitive graphs associated with \\hbox{\\rm PSL}(2,p), Discrete Math. 109 (1992), 161–170.", "source_ref_id": "dc31cf3db22a1b23ec4...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.0013263262808322906, -0.0030226127710193396, -0.02505050040781498, -0.013074468821287155, 0.021465318277478218, -0.061390507966279984, 0.050985854119062424, 0.023494379594922066, -0.0010002274066209793, -0.023616427555680275, -0.046073392033576965, -0.02096186764538288, 0.01236506085842847...
2db2de35f69e72f11e3f1649448b99e6ce0b5b37
subsection
41
77
Actions of
If ab\ne 0 and g^{\prime }\in gH then either \chi (g^{\prime })=(\xi ,y\eta ) or \chi (g^{\prime })=(1-\xi ,-y\eta ^{-1}) for some y\in S^*.Let \approx be the equivalence relation on F \times F^* defined by(\xi ,\eta ) \approx (1-\xi ,\frac{\xi \eta }{\xi - 1}) \textrm { for } \xi \ne 0,1.There is then a natural identi...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.02640535868704319, -0.019144266843795776, -0.015712028369307518, 0.0180764589458704, 0.006677611730992794, -0.056624315679073334, 0.004564878065139055, 0.00853483472019434, 0.04429876431822777, 0.026969771832227707, -0.05564803630113602, -0.006486931815743446, 0.009488234296441078, 0.02...
c75f50c710c9bacc6b4f641d2d69423c7655994d
subsection
42
77
Actions of
Finally, all remaining cosets in \mathcal {H} contain matrices with no entry equal to zero, where we have to bring in the equivalence relation \approx on characters defined by (REF ).For each \xi \in F define the following subsets of \mathcal {H} (in fact subsets of (F\times F^*)/_{\approx } \cup \lbrace \infty \rbrace...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 1327, "openalex_id": "", "raw": "D. Marušič and R. Scapellato, A class of non-Cayley vertex-transitive graphs associated with \\hbox{\\rm PSL}(2,p), Discrete Math. 109 (1992), 161–170.", "source_ref_id": "dc31cf3db22a1b23ec4...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.011889902874827385, 0.01062307134270668, -0.01976562850177288, 0.028969237580895424, -0.0034513534046709538, -0.015980396419763565, 0.041698604822158813, 0.02986975573003292, -0.005731268785893917, -0.01462961733341217, -0.03718074783682823, -0.005254298914223909, 0.015659872442483902, ...
2f6d6d54047f7d423d796282807e6e0bb8308204
subsection
43
77
Actions of
2 suborbits of length \frac{p-1}{4}, namely {\cal S}_{\frac{1}{2}}^+ and {\cal S}_{\frac{1}{2}}^- which are self-paired if and only if p\equiv {1\pmod {8}}.Example 7.3 The smallest admissible pair of primes p = 13 and q = 7 gives rise to the action of G=\hbox{\rm PSL}(2,13) on cosets of H=D_{12} with the following sub...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 1393, "openalex_id": "", "raw": "C. E. Praeger and M. Y. Xu, Vertex primitive transitive graphs of order a product of two distinct primes, J. Combin. Theory Ser. B 59 (1993), 245–266.", "source_ref_id": "5dfe58ee8c60a5542d53...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.02534826099872589, 0.04352392256259918, -0.03864045441150665, 0.0005288835964165628, 0.011499045416712761, 0.007546486798673868, 0.05359607934951782, -0.0081111378967762, 0.00828663818538189, -0.019564401358366013, -0.03726697713136673, -0.01378054078668356, 0.025806086137890816, -0.005...
cf8314aad7cf0728fa0c1d66dd4fa01b1cb039b0
subsection
44
77
Actions of
Also, the basic orbital graphs arising from Rows 8 and 9 of Table REF are isomorphic, that is, X(G,H,\mathcal {S}_\xi ^+) \cong X(G,H,\mathcal {S}_\xi ^-), where \xi \in S^*\cap S^*+1 and \xi \ne \frac{1}{2},1.With the explicit description of the suborbits of G on {\cal H} the construction of the corresponding general...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.020017879083752632, -0.01058872602880001, -0.020902816206216812, -0.00544312410056591, 0.020063651725649834, -0.03902876377105713, 0.007849999703466892, -0.005824562162160873, 0.000517325650434941, 0.0055499267764389515, -0.027509327977895737, -0.04531486704945564, 0.028577353805303574, ...
838c4c6f2b3e69c871017c36df9924e77c2ac9f5
subsection
45
77
Actions of
These areV_\infty &=&\lbrace H \left[ \begin{array}{cc} 1 & a \\ 0 & 1 \end{array} \right] \colon a\in F\rbrace \ \textrm { and }\\ V_x &=&\lbrace H \left[ \begin{array}{cc} 1 & x \\ -x^{-1} & 0 \end{array} \right] \left[ \begin{array}{cc} 1 & a \\ 0 & 1 \end{array} \right] \colon a\in F\rbrace = V_{-x}, \ x\in F^*...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.030672967433929443, 0.015588276088237762, -0.03561726585030556, 0.015992671251296997, -0.010865249671041965, -0.023531202226877213, 0.038303058594465256, 0.04907674714922905, 0.03360292315483093, 0.012688841670751572, -0.04828321933746338, -0.029833657667040825, 0.05020600184798241, 0.0...
b9303a08bbaffc0991de208cfca742a938e29f8d
subsection
46
77
Actions of
Since, by Theorem REF , \mathcal {S}_0 ^+ \cup \mathcal {S}_1 ^+ and \mathcal {S}_0 ^- \cup \mathcal {S}_1 ^- are the only two suborbits with nontrivial intersection with \mathcal {S}_1 we obtain the value of d(V_\infty ) as given in Table REF .To determine the values for d(V_\infty , V_x) let us consider the neighbors...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.4134/jkms.2013.50.6.1199", "end": 1962, "openalex_id": "https://openalex.org/W2068452045", "raw": "K. Kutnar, A. Malnič, L. Martínez and D. Marušič, Quasi m-Cayley strongly regular graphs, J. Korean Math. Soc. 50 (2013), 1199–1211.", ...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.045643337070941925, 0.019999470561742783, -0.03941924870014191, 0.02652866579592228, 0.009854812175035477, -0.009961597621440887, 0.06443765759468079, 0.0009052964160218835, 0.023859018459916115, 0.019114673137664795, -0.022821668535470963, -0.038900572806596756, 0.02713887020945549, 0....
aa9434a0d44ebcf7a4ded708d8cf9b2341346fd4
subsection
47
77
Actions of
Equivalently, g fixes a vertex and the only power g^i fixing another vertex is the identity mapping. If a group G is quasi-semiregular on the vertex set of the graph with m+1 orbits, then the graph is called a quasi m-Cayley graph on G. If G is cyclic and quasi-semiregular with two nontrivial orbits then the graph is s...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.045147836208343506, 0.006250182166695595, -0.025107569992542267, 0.024848099797964096, -0.002863712841644883, -0.04581940546631813, 0.04291944578289986, 0.027000175788998604, 0.01569032296538353, 0.020894993096590042, -0.009432509541511536, -0.04899410158395767, 0.02739701420068741, 0.0...
cc8ca3147dc79e3120203ffb64104bf2b7be507a
subsection
48
77
Actions of
Applying the action of on any of these elements gives \begin{eqnarray*} H \left[ \begin{array}{cc} 1 & x \\ -x^{-1} & 0 \end{array} \right] \left[ \begin{array}{cc} 1 & a \\ 0 & 1 \end{array} \right] \sigma &=& H \left[ \begin{array}{cc} 1 & x \\ -x^{-1} & 0 \end{array} \right] \sigma \left[ \begin{array}{cc} 1 & az^...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.002580819418653846, 0.036287885159254074, -0.043704163283109665, 0.06164972856640816, -0.006164973136037588, -0.02226409874856472, 0.07367447763681412, 0.004513096064329147, 0.028078094124794006, -0.001440146821551025, -0.00547828059643507, -0.014466322958469391, 0.01022408902645111, -0...
b3927189c674cbccf6ed33efa2f9d42ce15760cf
subsection
49
77
Actions of
If this neighbor is inside an orbit Vx, xF*, then there exists jF such that \left[ \begin{array}{cc} 1-\eta y^{-1} & y \\ \frac{\xi -1}{\eta }-\xi y^{-1} & \frac{y(\xi -1)}{\eta } \end{array} \right] \equiv \left[ \begin{array}{cc} 1 & x \\ -x^{-1} & 0 \end{array} \right] \left[ \begin{array}{cc} 1 & j \\ 0 & 1 \en...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.004954067058861256, 0.06626779586076736, -0.03270675614476204, 0.01003780122846365, 0.012051464058458805, 0.016872048377990723, 0.01341678760945797, -0.012448093853890896, 0.029259122908115387, -0.004366748966276646, -0.019465401768684387, -0.005324000958353281, -0.0016780518926680088, ...
cfb398b611c3606dc0a8469f0a08bde251c132b0
subsection
50
77
Actions of
Then for y \in F^* we haved(V_y)=\left\lbrace \begin{array}{cl} 0, & \xi \in N^*\cap N^*+1 \\ 2, & \xi \in S^*\cap N^*+1 \textrm { or } \xi \in N^*\cap S^*+1\\ 4, & \xi \in S^*\cap S^*+1 \\ 0, & \xi =1/2 \textrm { and } p\equiv {5\pmod {8}}\\ 2, & \xi =1/2 \textrm { and } p\equiv {1\pmod {8}}\\ \end{array}\right.,wher...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.019027164205908775, -0.0005659894668497145, -0.030577734112739563, 0.02708357200026512, 0.01776072010397911, -0.009666165336966515, -0.008163217455148697, -0.003850826295092702, 0.006042306777089834, 0.01600600965321064, -0.009872152470052242, -0.011703155003488064, -0.030867643654346466,...
3c392ca4457bd792d30f04ad895b990682f6cb79
subsection
51
77
Case
Let\varepsilon =\left\lbrace \begin{array}{ll} 2, & q\equiv {1,3\pmod {8}}\\ 0, & q\equiv {5,7\pmod {8}} \end{array}\right..The following lemma gives us the number of edges inside a block and between two blocks from \mathcal {L}_P for the orbital graph X(G,H,\mathcal {S}_0).Lemma 6.6 Let X=X(G,H,\mathcal {S}_0). Then ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.019682444632053375, 0.02493109554052353, -0.02088780328631401, -0.018370281904935837, 0.004783291835337877, -0.027494391426444054, 0.01096266321837902, -0.021269245073199272, 0.017714200541377068, -0.01296905241906643, -0.027860576286911964, -0.0005888522136956453, 0.009925139136612415, ...
928e00a19fabee17d293423fe9ff5a5dea8b060e
subsection
52
77
Case
This implies that d(\langle \mathbf {x}\rangle P,\langle \mathbf {y}\rangle P)=1 for \frac{q+1}{2} blocks \langle \mathbf {y}\rangle P\in \mathcal {L}_P, proving part (ii).To prove part (i), take \mathbf {y}=\pm \mathbf {x}=\pm (1,1,0,0). Then, by (REF ), there are edges inside the block \langle \mathbf {x}\rangle P if...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.033787235617637634, 0.011582886800169945, -0.03351254016160965, 0.01802290976047516, -0.003187582828104496, -0.053504083305597305, 0.03653416410088539, 0.022326434031128883, -0.00009067013888852671, 0.009049607440829277, -0.01892329379916191, -0.014520575292408466, 0.014581617899239063, ...
d768339aa5294a7f6152df449cadf0bb1c1f4eb6
subsection
53
77
Case
By Lemma REF , X\langle \mathcal {L} \rangle _\mathcal {P} is vertex-transitive, and consequently one can see that also X\langle \mathcal {L}\rangle ^{\prime }_\mathcal {P} is vertex-transitive.If q\equiv {1\pmod {4}} then Proposition REF and Lemma REF (iii) combined together imply that X\langle \mathcal {L}\rangle ^{\...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.030481282621622086, 0.023722920566797256, -0.03530214726924896, 0.018688473850488663, 0.028040340170264244, -0.020885324105620384, 0.03691927343606949, 0.050924185663461685, 0.000274844525847584, -0.010389266535639763, -0.027750477194786072, -0.007147388067096472, 0.008589070290327072, ...
571f28a2790df6c0fdd5b0cbf4ca02ef5c02a0c8
subsection
54
77
Case
Take a respective Hamilton path for each component, say \mathcal {U}=U_1U_2\cdots U_l, and \mathcal {U}^{\prime }=U^{\prime }_1 U^{\prime }_2\cdots , U^{\prime }_l, where l=\frac{q(q-1)}{4}. Choose any two isolated vertices W_1 and W_2 and construct the cycle \mathcal {D}=W_1\mathcal {U}W_2\mathcal {U}^{\prime }W_1. Ch...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.030860047787427902, 0.027809128165245056, -0.015353750437498093, 0.017664821818470955, 0.014514748007059097, -0.030814282596111298, 0.010441770777106285, 0.0308447927236557, 0.024163279682397842, 0.007894253358244896, -0.01778685860335827, 0.04460443928837776, 0.0009000211139209569, 0.0...
d1f8dc4132352c9a09e50dc9343c2aa5008d183f
subsection
55
77
Case
We can therefore conclude that X\langle \mathcal {L}\rangle _\mathcal {P}, which is vertex-transitive by Lemma REF , has at most two connected components. The rest of the argument follows word by word from the argument given in the proof of Proposition REF , since, by Lemma REF , d(\langle \mathbf {x}\rangle P, \langle...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.004822519142180681, 0.010865929536521435, -0.03433755785226822, 0.024433080106973648, -0.006470722146332264, -0.021289285272359848, 0.050758540630340576, 0.036138370633125305, 0.015001697465777397, 0.01016391720622778, -0.04035044461488724, 0.00843177828937769, 0.012735418975353241, 0.0...
3101250c1cdcb9c3cbbb53a6837b816ca80ba3c9
subsection
56
77
Case
Consequently, vertices inside each of these two sets are of the same valency.)Since we are in the case \mathcal {S}_\xi =\mathcal {S}_\xi ^+\cup \mathcal {S}_\xi ^- the solutions of () and () for x=y depends solely on \xi . Consequently, d(V_x)=d(V_y) for all x,y\in F^*. Suppose first that d(V_x)=0, x\in F^*. Note that...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.01434385683387518, -0.021225856617093086, -0.0406511016190052, 0.03173959627747536, 0.023484250530600548, -0.025254342705011368, 0.035188227891922, 0.053957317024469376, 0.03552393615245819, 0.002679936122149229, -0.019730431959033012, -0.0013743017334491014, 0.02217194065451622, 0.0017...
6c1122a8d17e04f8a87df13646fcc5a0dacf158f
subsection
57
77
Case
By Proposition REF we have d(V_\infty )=0 and d(V_\infty ,V_y)=1 for every y\in F^*.We now need to compute the valency val_{X_\mathcal {P}-\lbrace V_\infty \rbrace }(V_y), y\in F^*. Let x\in F^*. The number of edges d(V_y,V_x) between V_y and V_x in X_\rho is obtained from () and () by letting \xi =1/2. We obtainj_{1,...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.03234866261482239, -0.02059938572347164, -0.01267243642359972, 0.015426650643348694, 0.011581432074308395, -0.006031041964888573, 0.02719118818640709, 0.05029301717877388, 0.01051331590861082, 0.01116944383829832, -0.034942660480737686, 0.016479507088661194, -0.00048160599544644356, -0....
ff0edf9ca16c392acf32e22f3218b5c94446be40
subsection
58
77
Cases
Proposition 7.9 Let X=X(G,H,\mathcal {W}), where \mathcal {W}\in \lbrace \mathcal {S}_0^+\cup \mathcal {S}_1^+, \mathcal {S}_0^-\cup \mathcal {S}_1^-), be one of the graphs arising from Rows 3 or 4 of Table REF . Then X is hamiltonian.Proof. Suppose that \mathcal {W}=\mathcal {S}_0^+\cup \mathcal {S}_1^+. By Propositi...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.021604357287287712, 0.023969240486621857, -0.011618444696068764, 0.013701068237423897, 0.00619828375056386, -0.04204915836453438, 0.008772955276072025, 0.01827826164662838, 0.037197332829236984, 0.02053634449839592, -0.007682057563215494, 0.011549786664545536, -0.006961149629205465, 0.0...
6e4f1d407803317a5915f9e630485871887d25a8
subsection
59
77
Cases
Let \eta _i be the possible solution for \eta given in the i-th row of Table REF . Then \eta _1\eta _4=\eta _2\eta _3=y^2. This implies that either (\eta _1,\eta _4)\in S^*\times S^* or (\eta _1,\eta _4)\in N^*\times N^*, and that either (\eta _2,\eta _3)\in S^*\times S^* or (\eta _2,\eta _3)\in N^*\times N^*. [Table: ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.016148090362548828, 0.04694851115345955, -0.018071210011839867, 0.003998865373432636, 0.003760383464396, 0.007486425805836916, -0.01166844554245472, 0.04719271883368492, 0.048627424985170364, 0.009195864200592041, 0.01611756533384323, 0.013805244117975235, 0.003869131440296769, 0.0026061...
39ed5a3dc62ebd297c98a58bce89c58a041a8c09
subsection
60
77
Cases
(Namely, the number of possible number-theoretic solutions in Propositions REF and REF needs to be divided by 2 because of the fact that the equivalence relation (REF ) implies V_y=V_{-y}.) Moreover, (p-5)/8 is not an integer if p\equiv {1\pmod {8}} and so there are at least (p-1)/8 vertices in \mathcal {O}(S^*) adjac...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.002886335365474224, 0.001968736294656992, -0.035101499408483505, 0.02744021639227867, -0.004406764172017574, 0.010164173319935799, 0.04013780131936073, 0.027913322672247887, 0.02798963151872158, 0.016070382669568062, -0.029424212872982025, -0.015169952996075153, 0.009820789098739624, -0...
bcc022cd2dd539f9fd71ceed26aac6fffbdc99a5
subsection
61
77
Cases
Thus, since \pm 1\in N^*-1\cap S^* for p\equiv {5\pmod {8}}, Propositions REF and REF combined together imply thatval(V_x)_{X_\mathcal {P}\langle \mathcal {O}(N^*)\rangle }= \left\lbrace \begin{array}{cr} (p+7)/8, & p\equiv {1\pmod {8}}\\ (p-5)/8, & p\equiv {5\pmod {8}} \end{array}\right..If follows that, with the ex...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.0092922393232584, -0.017531663179397583, -0.023787522688508034, 0.019606776535511017, -0.002645389176905155, -0.03899994120001793, 0.05056564882397652, 0.06136845052242279, 0.016433073207736015, 0.002469919854775071, -0.02043071947991848, -0.0030859694816172123, -0.011077447794377804, -...
ec4c2efd3e4ae985cbf602f340157495f8bd5339
subsection
62
77
Cases
We obtainj_{1,2}=j_{3,4}=\frac{1}{2}(y-x\pm \sqrt{x^2+y^2}) \textrm { and } \eta _{1,2}=\eta _{3,4}=\frac{y}{x}(\pm \sqrt{x^2+y^2}-y).By Proposition REF , 2\in S^*, and so d(V_x)=2 for every x\in F^* for which \eta _{1,2}=\eta _{3,4}= \pm x(\sqrt{2}-1)\in S^*. Further, if x^2+y^2=0 then y^2=-x^2, and so y=\pm \sqrt{-1}...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.008262190967798233, -0.0024088521022349596, -0.044949982315301895, 0.011626573279500008, 0.010672950185835361, -0.03387269377708435, 0.01261071301996708, 0.022413957864046097, -0.00485584931448102, -0.002553802914917469, -0.025877518579363823, 0.0248857494443655, -0.003764904337003827, ...
ea9297ad81f2037e149e984c9d65cd9c4f763ee0
subsection
63
77
Cases
The corresponding values \eta _1 and \eta _2 for the pairs V_1,V_s and V_g, V_{gs} are, respectively,\eta _{1,2}(1,s)=s(\pm \sqrt{1+s^2} - s) \textrm { and } \eta _{1,2}(g,gs)=gs(\pm \sqrt{1+s^2} - s).Therefore\frac{\eta _{i}(g,gs)}{\eta _{i}(1,s)}=g\in N^*, i\in \lbrace 1,2\rbrace .And consequently, the bicirculant X_...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.045219872146844864, 0.009420806542038918, -0.024776339530944824, 0.010023432783782482, 0.02607312984764576, -0.01725495047867298, 0.01754482090473175, 0.026576589792966843, 0.04219910874962807, 0.005297773517668247, 0.0015704522375017405, -0.0012090670643374324, 0.02044353075325489, 0.0...
f7157553106742beca7171d0d496d9d4b9b75780
subsection
64
77
Cases
Of course, this cycle contains an edge of the form V_xV_y, x,y\in S^*. Replacing this edge with a path V_xV_\infty V_y gives a Hamilton cycle in X_\rho . Obviously this Hamilton cycle contains double edges in X_\rho , and so, by Lemma REF , lifts to a Hamilton cycle in X. \rule {2.5mm}{3mm}In Proposition REF graphs ari...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.039005592465400696, -0.019380712881684303, -0.026492061093449593, 0.013948008418083191, 0.014825481921434402, -0.011071423068642616, 0.003567119361832738, 0.03020033799111843, 0.006485671736299992, -0.012116760015487671, -0.01922810822725296, 0.03247413784265518, -0.03253518044948578, 0...
74cc2ddfe6b702f8cb7e80f53595f7e5357bb744
subsection
65
77
Cases
(For example, this can be checked using Magma .) The below matrix gives the symbol of this graph with respect to the orbits S_i=\lbrace v_i^j\colon j\in \mathbb {Z}_{13}\rbrace , i\in \mathbb {Z}_7, of a (7,13)-semiregular automorphism:\left[ \begin{array}{ccccccc} \emptyset & \lbrace 0,2\rbrace & \lbrace 6,12\rbrace &...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1006/jsco.1996.0125", "end": 48, "openalex_id": "https://openalex.org/W1976677460", "raw": "W. Bosma, J. Cannon, and C. Playoust, The Magma Algebra System I: The User Language, J. Symbolic Comput. 24 (1997), 235–265.", "source_ref_...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.03095536306500435, -0.002343774074688554, -0.03927013650536537, 0.010389651171863079, -0.008978428319096565, -0.033289603888988495, 0.03098587691783905, 0.03368626907467842, 0.015805695205926895, 0.0045120990835130215, -0.013761329464614391, -0.0174991637468338, -0.01462331973016262, 0....
5c6441082865eb73a8b57809fb57b94df5a04b28
subsection
66
77
Cases
Because of the isomorphism given in Proposition REF we may assume that X arises from Row 8 of Table REF , that is, it is associated with a suborbit \mathcal {S}_\xi ^+, where \xi \in S^*\cap S^*+1, \xi \ne \frac{1}{2} and \xi \ne 1.There are 8 self-paired suborbits of length (p-1)/2=18, giving 4 non-isomorphic vertex-...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1006/jsco.1996.0125", "end": 429, "openalex_id": "https://openalex.org/W1976677460", "raw": "W. Bosma, J. Cannon, and C. Playoust, The Magma Algebra System I: The User Language, J. Symbolic Comput. 24 (1997), 235–265.", "source_ref...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.00072409154381603, -0.02089807391166687, -0.0375555157661438, 0.009960906580090523, -0.010761745274066925, 0.004995707422494888, 0.026481062173843384, 0.01824386604130268, 0.0144761111587286, -0.005281721241772175, -0.02465057373046875, -0.016459139063954353, -0.030248817056417465, 0.022...
a2a10abd5bde900d0a4874f9584f7cdab52812ac
subsection
67
77
Cases
It follows by () and () that vertices V_1 and V_x, x\in F_{37}^*, are adjacent in X_\mathcal {P} if and only ifx^2+2(1-2\xi )x+1 \in S^*\cup \lbrace 0\rbrace \textrm { and } \eta _{1,2}(1,x)=1 - \frac{2}{1+x\pm \sqrt{x^2+2(1-2\xi )x+1}} \in S^*orx^2-2(1-2\xi )x+1 \in S^*\cup \lbrace 0\rbrace \textrm { and } \eta _{3,4}...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.02449876442551613, 0.03325485810637474, -0.024819109588861465, 0.020120715722441673, -0.011692591942846775, 0.0013681399868801236, 0.04860090836882591, 0.043597426265478134, 0.0441771000623703, 0.022302111610770226, -0.012020563706755638, -0.05378744751214981, -0.007127675693482161, 0.0...
90aff9c414fefe958bbebbee4f6382a59f2e775d
subsection
68
77
Cases
Therefore, connectedness of X implies that the bipartite graph induced by the edges with one endvertex in \mathcal {O}(S^*) and the other in \mathcal {O}(N^*) contains a matching preserved by the \sigma . Since, by (), we have\frac{\eta _{i}(\tau ,\tau ^3)}{\eta _{i}(1,\tau ^2)}=\tau \in N^*, i\in \lbrace 1,2,3,4\rbrac...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.03598511964082718, 0.004929259419441223, -0.020617429167032242, 0.022525038570165634, 0.008423998020589352, -0.020083298906683922, 0.038030076771974564, 0.040868598967790604, 0.032810863107442856, 0.02156360261142254, 0.004128064028918743, -0.005303150974214077, 0.0073328460566699505, 0...
0eb9213edb0eaf68d59d2f170d8c3602b4a60aa3
subsection
69
77
Cases
As before this cycle/path can be extended to a Hamilton cycle in X_\rho which, by Lemma REF , lifts to a Hamilton cycle in X. [Table: Information about existence of edges inX(\hbox{\rm PSL}(2,37),D_{36},\mathcal {S}_\xi ^+) where \xi \in S^*\cap S^*+1, \xi \ne \frac{1}{2} and \xi \ne 1.]Example 7.14 Let X be a basic ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1006/jsco.1996.0125", "end": 872, "openalex_id": "https://openalex.org/W1976677460", "raw": "W. Bosma, J. Cannon, and C. Playoust, The Magma Algebra System I: The User Language, J. Symbolic Comput. 24 (1997), 235–265.", "source_ref...
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.034787196666002274, -0.021269015967845917, -0.023817025125026703, 0.02461041696369648, 0.005687249358743429, -0.022169210016727448, 0.026334518566727638, 0.03374968469142914, 0.03083549439907074, -0.017347825691103935, -0.010764196515083313, -0.013731788843870163, -0.014075083658099174, ...
a9782329dff3346ad9758ae23a252801f0aec897
subsection
70
77
Cases
Next, by (), we have\frac{\eta _{i}(\tau ,\tau ^3)}{\eta _{i}(1,\tau ^2)}=\tau \in N^*, i\in \lbrace 1,2,3,4\rbrace ,and so Table REF implies that vertices in X\langle \mathcal {O}(N^*)\rangle are of valency at least 2. Namely, for every \xi there exists at least one \tau \in \mathcal {T}^+\cup \mathcal {T}^- such tha...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.033508460968732834, -0.0006904635229147971, -0.0122528662905097, 0.003835696494206786, 0.011711176484823227, -0.0126724848523736, 0.029052872210741043, 0.03210464492440224, 0.04074116051197052, 0.005790737923234701, 0.001814850838854909, -0.005943326745182276, 0.011047416366636753, 0.01...
b397ce572c525dbe61a934d2ae63f0a6c01e316e
subsection
71
77
Cases
The number of edges d(V_y,V_x), x,y\in F^*, between V_y and V_x in X_\rho is obtained from () and ():j_{1,2}&=&\frac{1}{2}(y-x\pm \sqrt{x^2+2(1-2\xi )xy+y^2}), \\ j_{3,4}&=&\frac{1}{2}(y-x\pm \sqrt{x^2-2(1-2\xi )xy+y^2}), \\ \eta _{1,2}&=&\frac{y}{2\xi x}((2\xi -1)x-y \pm \sqrt{x^2+2(1-2\xi )xy+y^2}),\\ \eta _{3,4}...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.009406649507582188, -0.03518529608845711, -0.01791306957602501, -0.011176595464348793, 0.012809217907488346, -0.03222521394491196, -0.003160344436764717, 0.029234616085886955, -0.010200073942542076, -0.00958974752575159, -0.04165475070476532, 0.008956534788012505, -0.017607904970645905, ...
9519f8df2c6b14fc7e9992d844130e20c388e5f9
subsection
72
77
Cases
Consequently, Theorem REF and Proposition REF combined together imply that for the two polynomialsf(z)=1+2(1-2\xi )z^2+z^4 \textrm { and } h(z)=1-2(1-2\xi )z^2+z^4there exist g,g^{\prime }\in F^* such that F^*=\langle g\rangle =\langle g^{\prime }\rangle and f(g),h(g^{\prime })\in S^*\cup \lbrace 0\rbrace . It follows ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ 0.0014009964652359486, 0.01956244930624962, -0.022431202232837677, -0.01001774426549673, 0.008888554759323597, -0.026139216497540474, 0.03399776667356491, 0.0138478334993124, 0.008438404649496078, -0.0014086260925978422, -0.04226832091808319, 0.0060083577409386635, -0.032349761575460434, -...
c2da4c9772177b762fa3b5b8fced5cf9efca8b76
subsection
73
77
Cases
Hence,\frac{\eta _{i}(\bar{g},\bar{g}\bar{s})}{\eta _{i}(1,\bar{s})}=\bar{g}\in N^*, i\in \lbrace 1,2,3,4\rbrace .Consequently, for each i exactly one of \eta _{i}(\bar{g},\bar{g}\bar{s}) and \eta _{i}(1,\bar{s}) belongs to S^*, implying that in the bicirculant X_\rho -\lbrace V_\infty \rbrace we have a full induced c...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.022614141926169395, 0.036072760820388794, -0.01599164679646492, 0.018219491466879845, 0.016403645277023315, -0.024552062153816223, 0.025253985077142715, 0.011856403201818466, 0.02137814648449421, 0.01249728910624981, 0.009537003934383392, 0.006195237394422293, 0.003444765694439411, 0.01...
5197d940a34e7b084d8b5315aabd357434714fee
subsection
74
77
Cases
This implies thatval(X)=\frac{p-1}{2}&\ge & 4 + d(V_x) + |\sum d(V_x,V_y) \colon y\in S^*|.On the other hand, since, by assumption the valency of the graph induced on \mathcal {O}(S^*) is either 0 or 1, we have by calculating valency val(X) at V_x, x\in S^*:val(X)=\frac{p-1}{2}&=&d(V_\infty ,V_1) + \epsilon + d(V_1) + ...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.01811448112130165, 0.021654969081282616, -0.02917850762605667, 0.014657926745712757, 0.01336076483130455, -0.013528632931411266, 0.02614162303507328, 0.026660487055778503, 0.020083114504814148, -0.004120395984500647, -0.03131501004099846, 0.006894032936543226, 0.03665626421570778, -0.01...
52bfdb047565ac18f72680cd8fb3c62e836d46bb
subsection
75
77
Cases
Replacing this edge with a path V_xV_\infty V_y gives a Hamilton cycle in X_\rho . Obviously this Hamilton cycle contains double edges, and so, by Lemma REF , it lifts to a Hamilton cycle in X. We may therefore assume that (p-1)/4\equiv {5\pmod {6}} and that the generalized Petersen graph in X_\mathcal {P}-\lbrace V_\i...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.05849793553352356, -0.006946058012545109, -0.008414609357714653, 0.017683647572994232, 0.01664612628519535, -0.025053106248378754, 0.003061071503907442, 0.03957841545343399, 0.026334751397371292, -0.009360585361719131, -0.033261738717556, 0.013602219521999359, -0.013709023594856262, -0....
fe1fd748c339b1a835470129c8016a401e959c20
subsection
76
77
Proof of Theorem
Proof of Theorem REF . Let X be a connected vertex-transitive graph of order pq, where p and q are primes and p\ge q, other than the Petersen graph. If q\in \lbrace 2,p\rbrace then X admits a Hamilton cycle by Proposition REF . We may therefore assume that q\notin \lbrace 2,p\rbrace . Then X is a generalized orbital gr...
{ "cite_spans": [] }
1808.08553
Hamilton cycles in vertex-transitive graphs of order a product of two primes
[ "Shaofei Du", "Klavdija Kutnar", "Dragan Marusic" ]
[ "math.CO" ]
2,018
en
Mathematics
[ -0.006717944052070379, -0.0024567609652876854, -0.022492332383990288, 0.014153079129755497, 0.02787889540195465, -0.042573682963848114, -0.0018110981909558177, 0.02842823415994644, 0.009941489435732365, -0.0042802575044333935, -0.011597132310271263, -0.01786874048411846, -0.01976090297102928...
0440db2af16208ef9627ca40512cebb3b9451539
abstract
0
53
Abstract
In this paper, we present new stochastic methods for solving two important classes of nonconvex optimization problems. We first introduce a randomized accelerated proximal gradient (RapGrad) method for solving a class of nonconvex optimization problems consisting of the sum of $m$ component functions, and show that it ...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.026817237958312035, 0.015338300727307796, 0.009465355426073074, -0.03261391073465347, -0.02516976185142994, 0.014194220304489136, 0.04750220477581024, 0.008984842337667942, 0.0031957970932126045, 0.0538785457611084, -0.05448872223496437, 0.0027343514375388622, -0.009785698726773262, 0.0...
7cc1a0ad9f945c2b03ac583f73b81c6e4ed922af
subsection
1
53
Introduction
Nonconvex optimization plays a fundamental role in modern statistics and machine learning, e.g., for empirical risk minimization with either nonconvex loss () or regularization (, , ), as well as the training of deep neural networks (). In this paper, we consider two classes of nonconvex optimization problems that are ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 236, "openalex_id": "", "raw": "Robert Tibshirani, G James, D Witten, and T Hastie, An introduction to statistical learning-with applications in r, 2013.", "source_ref_id": "43a1b3c47ecd0cf6895da5b2cb7783b6de63cd60", "...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.03169989213347435, -0.008878410793840885, -0.04475817456841469, -0.011441250331699848, 0.022119751200079918, 0.08036334812641144, 0.03206601366400719, 0.03502548113465309, 0.028252262622117996, 0.03771036118268967, -0.03307284042239189, -0.008321602828800678, -0.010777657851576805, -0.0...
08f46c39d989084b71443542865a2924a037af85
subsection
2
53
Introduction
It can be shown that the condition number for these problems is usually larger than m (see Section  for more details).In addition to (REF ), we consider an important class of nonconvex multi-block optimization problems with linearly coupled constraints, i.e.,\min _{x_i\in X_i} &~ \textstyle {\sum }_{i=1}^mf_i(x_i)\\ \t...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.01868884451687336, -0.021785853430628777, -0.029642794281244278, -0.018582051619887352, -0.010931067168712616, 0.04366324096918106, 0.011960860341787338, 0.031610846519470215, 0.023296216502785683, 0.0171174556016922, -0.04235120862722397, -0.009489355608820915, -0.027018729597330093, 0...
b5b702c3552ebfa42c894423619337330f7432e4
subsection
3
53
Introduction
\Vert \nabla f(\bar{x})\Vert ^2 \le \epsilon . Since each GD iteration requires a full gradient computation, i.e., m gradient computations for f_i's, totally this algorithm needs {\cal O}(m L / \epsilon ) gradient computations for all the component functions f_i's. Ghadimi and Lan  (see also  ) show that by using the s...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 688, "openalex_id": "", "raw": "Saeed Ghadimi and Guanghui Lan, Stochastic first-and zeroth-order methods for nonconvex stochastic programming, SIAM Journal on Optimization, 23 (2013), pp. 2341–2368.", "source_ref_id": "9520...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.02880600467324257, 0.00006901597225805745, -0.037899427115917206, -0.03960825502872467, 0.010054638609290123, 0.041103482246398926, 0.025510402396321297, -0.01252633985131979, 0.040249068289995193, 0.0007905249367468059, -0.06215871870517731, -0.004115688614547253, -0.015066700056195259, ...
1804f64f3d87a9d0ace4dd6c1e03bb8599ec692c
subsection
4
53
Introduction
For example, In , Hong et al. established the complexity for a variant of ADMM for nonconvex multi-block problems, see also and for some previous work on the asymptotic analysis of ADMM for nonconvex optimization. In , Melo and Monteiro presented a linearized proximal multiblock ADMM with complexity \mathcal {O}\left(1...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.48550/arxiv.1604.00543", "end": 215, "openalex_id": "https://openalex.org/W2341508215", "raw": "Mingyi Hong, Decomposing linearly constrained nonconvex problems by a proximal primal dual approach: Algorithms, convergence, and application...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.004344087094068527, -0.013371703214943409, -0.006369294133037329, -0.04967286437749863, -0.01283775083720684, 0.021739505231380463, 0.015637189149856567, -0.001481718965806067, 0.007543989922851324, 0.018551046028733253, -0.051381513476371765, 0.0015694397734478116, -0.038444600999355316, ...
fc87f2078d822986cfd07faa96155dacee944691
subsection
5
53
Introduction
In particular, RapGrad does not require the computation of full gradients throughout its entire procedure by properly initializing a few intertwined search points and gradients using information obtained from the previous subproblems. This comes with the price of requiring additional storage (memory) for maintaining \m...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.010742587968707085, -0.011223257519304752, -0.0038568025920540094, -0.03756853938102722, -0.034730296581983566, 0.015991806983947754, 0.06671391427516937, 0.010383992455899715, -0.013779199682176113, 0.014198832213878632, -0.06402826309204102, -0.009529468603432178, -0.031007014214992523,...
07e8520a83636c5ffb92a3af51a581fb0f4074ee
subsection
6
53
Introduction
Moreover, we demonstrate that the total number primal block updates that RapGrad requires can be much smaller, up to a factor of {\cal O}(\sqrt{m}), than its batch counterpart. To the best of our knowledge, this is the first time that the complexity of randomized methods for solving this special class of nonconvex mul...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.03163757175207138, 0.010106022469699383, 0.029242634773254395, -0.044725824147462845, -0.02587141841650009, 0.01656625047326088, 0.055190324783325195, -0.005922510754317045, 0.005823357030749321, 0.05247504636645317, -0.06620398163795471, 0.0018791481852531433, -0.03294944763183594, 0.0...
2531fc370a1053fb8b17e9abd43d7ffea4730459
subsection
7
53
Notation and terminology
Let \mathbb {R} denote the set of real numbers. All vectors are viewed as column vectors, and for a vector x \in \mathbb {R}^d, we use x^{\top } to denote its transpose. For any n \ge 1, the set of integers \lbrace 1,\ldots ,n\rbrace is denoted by [n]. We use \mathbb {E}_s[X] to denote the expectation of a random varia...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.04868575185537338, -0.008920634165406227, -0.029638787731528282, 0.024724425747990608, 0.0036514312960207462, -0.015094108879566193, 0.011034419760107994, 0.017566552385687828, 0.029516691341996193, 0.019245371222496033, -0.005204339046031237, 0.023808706551790237, 0.0005966484895907342, ...
1d0e5490dbbc765d6f047b83d7902297f556f73a
subsection
8
53
Nonconvex finite-sum optimization
In this section, we develop a randomized accelerated proximal gradient (RapGrad) method for solving the nonconvex finite-sum optimization problem in (REF ) and demonstrate that it can significantly improve the existing rates of convergence for solving these problems, especially when their objective functions have large...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1198/016214501753382273", "end": 627, "openalex_id": "https://openalex.org/W2074682976", "raw": "Jianqing Fan and Runze Li, Variable selection via nonconcave penalized likelihood and its oracle properties, Journal of the American statist...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.05075468495488167, 0.012444511987268925, 0.001066291006281972, -0.05725543573498726, -0.01554228737950325, 0.01989901065826416, 0.05221964418888092, -0.0015546102076768875, -0.0007701520225964487, 0.05679763853549957, -0.01799151301383972, 0.037112265825271606, -0.04236169904470444, 0.0...
75707fafb35355d580512fc0b5a9198a7908aea4
subsection
9
53
Nonconvex finite-sum optimization
Using p_{\lambda ,\gamma ,\epsilon }, our problem of interest is given by\min _{x\in \mathbb {R}^n} \tfrac{1}{2m}\Vert Ax-b\Vert ^2 + \tfrac{\rho }{2}\textstyle {\sum }_{i=1}^np_{\lambda ,\gamma ,\epsilon }(x_i),which can be viewed as a special case of problem (REF ) with f_i(x) = \tfrac{1}{2}(a_i^{\top }x-b_i)^2 + \tf...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 2047, "openalex_id": "https://openalex.org/W2962851402", "raw": "Zeyuan Allen-Zhu and Elad Hazan, Variance reduction for faster non-convex optimization, in International Conference on Machine Learning, 2016, pp. 699–707.", "...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.028523661196231842, -0.014506014063954353, -0.0429457388818264, -0.05475810542702675, -0.003937455825507641, -0.004395299591124058, 0.01578034646809101, 0.019565187394618988, -0.008538784459233284, 0.014948596246540546, -0.04044285789132118, 0.020725058391690254, -0.028615230694413185, ...
d0ebccb2294d22f445de209ad49b53aad050a23e
subsection
10
53
Nonconvex finite-sum optimization
Figure REF shows our Algorithm REF not only reduces the function value as well as gradient norm faster than both SVRG and AG. [Figure: Comparison on function value f and square of gradient norm \Vert \nabla f\Vert ^2 for Algorithm , SVRG and AG. Left Figure: Comparison on function value f. Right Figure: Comparison on s...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.03050805814564228, 0.001767490990459919, -0.048989929258823395, -0.02599060721695423, -0.028218809515237808, -0.0015700432704761624, 0.018649747595191002, -0.009881924837827682, 0.007230211514979601, 0.054667264223098755, -0.05759750306606293, 0.005013455171138048, -0.029302386566996574, ...
0cf4fd392de98321da5465240c939ab73995e278
subsection
11
53
The Algorithm
The basic idea of RapGrad is to solve problem (REF ) iteratively by using the proximal-point type method. More specifically, given a current search point \bar{x}^{\ell -1} at the l-th iteration, we will employ a randomized accelerated gradient (RaGrad) obtained by properly modifying the randomized primal-dual gradient ...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1007/s10107-017-1173-0", "end": 512, "openalex_id": "https://openalex.org/W2964037929", "raw": "Guanghui Lan and Yi Zhou, An optimal randomized incremental gradient method, Mathematical programming, (2017), pp. 1–49.", "source_ref_...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.03255762904882431, -0.020520156249403954, -0.008101265877485275, -0.04326777905225754, -0.0004977472708560526, 0.012327349744737148, 0.0682276114821434, 0.004866862203925848, -0.016110992059111595, 0.009466733783483505, -0.07060763984918594, -0.02030656300485134, -0.0289723239839077, 0....
4e8175a5cc90a371f30c7adb2840014a87eaab5f
subsection
12
53
The Algorithm
By using the strong convexity of the objective functions, we will be able to show that all the search points \underline{x}^i_s, i = 1, \ldots , m, will converge, similarly to the search point x^s, to the optimal solution of the subproblem in (REF ) (see Lemma REF below). Therefore, we can use y^i_s to approximate \nabl...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.0006117860320955515, -0.012359698303043842, -0.042846955358982086, -0.024490514770150185, -0.032379359006881714, 0.0382998064160347, 0.035583723336458206, -0.02632158063352108, 0.02322402596473694, 0.023452909663319588, -0.06561321765184402, 0.012573323212563992, -0.010360784828662872, ...
b066f2ec75837ba81a7a4f57b255182d0cca4359
subsection
13
53
The Algorithm
Assume nonnegative parameters \lbrace \alpha _t\rbrace , \lbrace \tau _t\rbrace , \lbrace \eta _t\rbrace are given. t = 1, \ldots , s 1. Generate a random variable i_t uniformly distributed over [m]. 2. Update x^t and y^t according to\tilde{x}^t =&~ \alpha _t (x^{t-1}-x^{t-2}) + x^{t-1}.\\ \underline{x}_i^t = &~ \left\...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.01781928725540638, -0.0013110830914229155, -0.02492869459092617, -0.0030092932283878326, 0.004275560844689608, -0.0007895103772170842, 0.002320855390280485, 0.03016158565878868, 0.016415713354945183, 0.02463882602751255, -0.05303068831562996, 0.0054197791032493114, -0.009954696521162987, ...
eacb88009bbe22e42157d8901b734504324f565a
subsection
14
53
The Algorithm
A point x \in X is called an approximate stationary point if it sits within a small neighborhood of a point \hat{x} \in X which approximately satisfies the first-order optimality condition.A point x \in X is called an (\epsilon ,\delta )-solution of (REF ) if there exists some \hat{x}\in X such that[d\left( \nabla f(\h...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.03161345049738884, -0.009238390251994133, -0.0443381704390049, -0.00899427104741335, -0.013243473134934902, 0.006801011972129345, 0.02772280015051365, 0.011961846612393856, 0.05102093890309334, 0.0019586761482059956, -0.0208569448441267, 0.02105529233813286, -0.008673864416778088, 0.0052...
11eaef66071cb684f8a3df4208c7211f145d5de6
subsection
15
53
The Algorithm
\Vert \nabla f(\hat{x})\Vert ^2\le \epsilon and \Vert x-\hat{x}\Vert ^2\le \epsilon /L^2, which implies that\Vert \nabla f(x)\Vert ^2&=\Vert \nabla f(x)-\nabla f(\hat{x})+\nabla f(\hat{x})\Vert ^2 \le 2 \Vert \nabla f(x)-\nabla f(\hat{x})\Vert ^2+2\Vert \nabla f(\hat{x})\Vert ^2\\ &\le 2 L^2\Vert x-\hat{x}\Vert ^2+2\Ve...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1137/17m1151031", "end": 1203, "openalex_id": "https://openalex.org/W2962970637", "raw": "Damek Davis and Benjamin Grimmer, Proximally guided stochastic subgradient method for nonsmooth, nonconvex problems, arXiv preprint arXiv: 1707.035...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.01479303278028965, 0.005905006546527147, -0.010810586623847485, -0.012977281585335732, -0.0196986123919487, -0.006709887646138668, 0.03460608422756195, 0.012847584672272205, -0.015327077358961105, 0.0020064814016222954, -0.03353799507021904, 0.0012874286621809006, -0.028960470110177994, ...
1295d9599e96ffb31e06fe038198d9d6699cc589
subsection
16
53
The Algorithm
Suppose that in Algorithm REF , the number of iterations s=\lceil -\log \widetilde{M} / \log \alpha \rceil with\widetilde{M}:=\textstyle {6\left(5+\tfrac{2L}{\mu }\right)}\max \left\lbrace \tfrac{6}{5},\tfrac{L^2}{\mu ^2}\right\rbrace , \quad \alpha =1-\tfrac{2}{m\left(\sqrt{1+16c/m}+1\right)}, \quad c = 2+\tfrac{L}{\m...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.027028050273656845, -0.010049672797322273, -0.022785363718867302, -0.02252591960132122, -0.012254954315721989, 0.006783720105886459, 0.006623474881052971, 0.007584946695715189, 0.009782597422599792, 0.010599086061120033, -0.0345824733376503, 0.027150141075253487, -0.029134131968021393, 0...
bd09d6a27e78725c6e93e45497588886336f09e9
subsection
17
53
The Algorithm
Indeed, observe that the full gradient is computed only once in the first outer loop, and that for each subproblem (REF ), we only need to compute s gradients with\textstyle {s= \left\lceil -\tfrac{\log \widetilde{M}}{\log {\alpha }}\right\rceil \sim \mathcal {O}\left(\left(m+\sqrt{m\tfrac{ L}{\mu }}\right)\log \left(\...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.48550/arxiv.1703.10993", "end": 1481, "openalex_id": "https://openalex.org/W2604676813", "raw": "Courtney Paquette, Hongzhou Lin, Dmitriy Drusvyatskiy, Julien Mairal, and Zaid Harchaoui, Catalyst acceleration for gradient-based non-conve...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.04027331620454788, 0.01856539025902748, -0.01696361042559147, -0.03713078051805496, -0.018199268728494644, -0.004294294863939285, 0.015506752766668797, 0.021768948063254356, -0.006109645590186119, 0.017924677580595016, -0.055406320840120316, -0.006502463016659021, 0.008603844791650772, ...
84691da186d6a723f69b06874cc3f982123ac727
subsection
18
53
The Algorithm
In fact, our complexity bound minorizes those for variance-reduced stochastic algorithms as long as L/\mu \log (L/\mu ) > m^{\frac{1}{3}}.Theorem REF only shows the convergence of RapGrad in expectation. Similarly to the nonconvex SGD methods in , , we can establish and then further improve the convergence of RapGrad w...
{ "cite_spans": [ { "arxiv_id": "", "doi": "", "end": 695, "openalex_id": "", "raw": "Saeed Ghadimi and Guanghui Lan, Stochastic first-and zeroth-order methods for nonconvex stochastic programming, SIAM Journal on Optimization, 23 (2013), pp. 2341–2368.", "source_ref_id": "9520...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.05834675207734108, -0.001713668811134994, 0.00840718112885952, -0.029478536918759346, -0.03277427330613136, 0.03854181244969368, 0.04299716278910637, 0.021910548210144043, -0.024977413937449455, 0.05837726593017578, -0.06847809255123138, -0.012870158068835735, 0.005836200900375843, 0.03...
7d6104d2009d78472c7a1c6cab9ca029477f9912
subsection
19
53
The Algorithm
Indeed, by the method of multipliers and Fenchel conjugate duality, we have&\min _{{\textbf {x}}\in X,x_m\in \mathbb {R}^{n}} \lbrace \psi ({\textbf {x}})+\psi _m(x_{m}) + \max _{y\in \mathbb {R}^{n}} \left\langle \textstyle {\sum }_{i=1}^{m}\mathbf {A}_ix_i -\mathbf {b}, y\right\rangle \rbrace \\ =&\min _{{\textbf {x...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.006122183986008167, 0.005420326255261898, -0.023176567628979683, -0.028104830533266068, -0.01907222531735897, 0.03878527507185936, 0.0311106126755476, 0.030546074733138084, 0.018080469220876694, 0.01980459876358509, -0.02778441645205021, 0.04256920516490936, -0.02743348851799965, 0.03039...
e3d91a64bfc7460d061da63a8fbf66e5a518dcf6
subsection
20
53
The Algorithm
Specifically,y^t &=\textstyle \operatornamewithlimits{arg\,min}\limits _{y \in \mathbb {R}^n}~ h(y)+\langle -\mathbf {A}\tilde{{\textbf {x}}}^t+\mathbf {b},y\rangle + \tau _t V_{ h}(y,y^{t-1})\\ &=\textstyle \operatornamewithlimits{arg\,max}\limits _{y \in \mathbb {R}^n}~ \langle (\mathbf {A}\tilde{{\textbf {x}}}^t-\ma...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.01073533482849598, -0.004711491987109184, -0.004104911349713802, -0.020967094227671623, 0.003338101552799344, 0.020646635442972183, 0.04849594086408615, 0.02078397385776043, 0.013588935136795044, 0.0267200730741024, -0.08118262887001038, -0.007614686619490385, 0.00163090112619102, 0.050...
4f1abd7aee3b6b634ee08009f255e39f616accb3
subsection
21
53
The Algorithm
We will derive the convergence result for Algorithm REF in terms of primal variables and construct relations between successive search points ({\textbf {x}}^{\ell },x_m^l), which will be used to prove the final convergence of RapDual. [H] RapDual for nonconvex multi-block optimization Compute A_m^{-1} and reformulate...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.00045171004603616893, -0.007639263290911913, -0.017437199130654335, -0.024256471544504166, -0.0399087630212307, 0.0692606195807457, 0.05086231231689453, 0.018749184906482697, 0.02372252382338047, 0.033531904220581055, -0.053730372339487076, -0.0045499955303967, -0.029519669711589813, 0.0...
627cfc1c6ac5b80f979700f64896c4ebaaf9b515
subsection
22
53
The Algorithm
Update x^t and y^t according to\tilde{{\textbf {x}}}^t =&~ \alpha _t ({\textbf {x}}^{t-1}-{\textbf {x}}^{t-2}) + {\textbf {x}}^{t-1},\\ g^t=&~(\tau _t g^{t-1}+\mathbf {A}\tilde{{\textbf {x}}}^t-\mathbf {b})/(1+\tau _t), \\ y^t =&~\textstyle \operatornamewithlimits{arg\,min}\limits _{y \in \mathbb {R}^n}~ h(y)+\langle -...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.0025661264080554247, -0.009012922644615173, -0.00799028854817152, -0.016514789313077927, 0.004121065605431795, 0.022131647914648056, 0.029824305325746536, 0.009310555644333363, 0.01976585201919079, 0.021826384589076042, -0.0460338294506073, -0.01189003698527813, -0.01292793545871973, 0.0...
e1b44e532e182ce0a1e981843c3eae6a89dadb9c
subsection
23
53
The Algorithm
A point (\emph {\textbf {x}},x_m) \in X\times \mathbb {R}^{n} is called an (\epsilon ,\delta ,\sigma )-solution of (REF ) if there exists some \hat{\emph {\textbf {x}}}\in X, and \lambda \in \mathbb {R}^n such that\left[d(\nabla f(\hat{\emph {\textbf {x}}})+ \mathbf {A}^{\top }\lambda , -N_{X} (\hat{\emph {\textbf {x}}...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.031438399106264114, -0.011835183016955853, -0.019687151536345482, -0.014505920931696892, -0.011049223132431507, -0.004353302530944347, 0.0360168032348156, 0.019076697528362274, 0.043678008019924164, -0.00863029807806015, 0.006260972004383802, 0.0287676602602005, -0.007581079378724098, 0....
b71cbd07102f54d85880abf4940c6c2863f1ab8a
subsection
24
53
The Algorithm
Besides, the definition of a (\epsilon ,\delta ,\sigma )-solution guarantees \Vert \nabla f_m(x_m)+\lambda \Vert ^2\le \epsilon and \Vert \mathbf {A}{\textbf {x}}+x_m-\mathbf {b}\Vert ^2\le \sigma , which altogether justify that ({\textbf {x}},x_m) is a reasonably good solution. Let the iterates (\emph {\textbf {x}}^...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.0065629188902676105, -0.0179335568100214, -0.013598978519439697, -0.006192800588905811, 0.0015300757950171828, 0.0064675272442400455, 0.018757738173007965, 0.01736884005367756, 0.021367643028497696, 0.005078630521893501, 0.003365403739735484, 0.01715516485273838, -0.012393232434988022, ...
0f04bd24c969510848a4320a71a2f6c424aa104e
subsection
25
53
The Algorithm
Suppose in Algorithm REF , number of iterations s=\left\lceil -\log \widehat{\mathcal {M}}/\log \alpha \right\rceil with\widehat{\mathcal {M}}=(2+\tfrac{L}{\mu })\cdot \max \left\lbrace 2,\tfrac{L^2}{\mu ^2}\right\rbrace , \quad \alpha =1-\tfrac{2}{(m-1)(\sqrt{1+8c}+1)},\quad c = \tfrac{\bar{A}^2}{\mu \bar{\mu }}=\tfra...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.013332574628293514, -0.023215623572468758, -0.03754794970154762, -0.014256009832024574, -0.015996037051081657, 0.017369743436574936, -0.005613113287836313, 0.013584421016275883, 0.019048716872930527, 0.018682394176721573, -0.020437685772776604, 0.02231508307158947, -0.022986670956015587, ...
8052ed4df3c4166f35bec1e5604586937b0b1a89
subsection
26
53
The Algorithm
According to Theorem REF , we can bound the complexity of RapDual to compute a stochastic (\epsilon ,\delta ,\sigma )-solution of (REF ) in terms of block updates in (). Note that for each subproblem (REF ), we only need to update s primal blocks with\textstyle {s= \left\lceil -\tfrac{\log \widehat{\mathcal {M}}}{\log ...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.04328089579939842, -0.0005641892203129828, 0.02000749483704567, -0.02870640531182289, -0.027027668431401253, 0.011835097335278988, 0.03397153690457344, 0.015017067082226276, 0.026127254590392113, 0.026890316978096962, -0.048805467784404755, 0.022617168724536896, -0.016161659732460976, 0...
366270a187e108640862e190b716f31e62f1c72c
subsection
27
53
The Algorithm
\mathbf {A}_1 = \mathbf {A}_2=\ldots =\mathbf {A}_{m-1}, we immediately have \Vert \mathbf {A}\Vert = \sqrt{m-1}\bar{A}, which means that RapDual can potentially save the number of primal block updates by a factor of {\cal O}(\sqrt{m}) than its batch counterpart.It is also interesting to compare RapDual with the noncon...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1137/130936361", "end": 372, "openalex_id": "https://openalex.org/W1975768153", "raw": "Cong D. Dang and Guanghui Lan, Stochastic block mirror descent methods for nonsmooth and stochastic optimization, SIAM Journal on Optimization, 25 (2...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.021130742505192757, 0.016660483554005623, -0.0028568459674715996, -0.041162990033626556, -0.0007280380232259631, 0.011152760125696659, 0.039637304842472076, 0.013509944081306458, 0.026668980717658997, 0.024533020332455635, -0.06377364695072174, 0.026119733229279518, -0.019025295972824097,...
23c5595e7224ecf2e7157e0bfc9612eda651882c
subsection
28
53
The Algorithm
To do so, we can iteratively solve the following saddle-point subproblems in place of the ones in (REF ):&\min _{{\textbf {x}}\in X,x_m\in \mathbb {R}^{n}} \lbrace \psi ({\textbf {x}})+\psi _m(x_{m}) + \max _{y\in \mathbb {R}^{n}} \left\langle \textstyle {\sum }_{i=1}^{m} A_ix_i - b , y\right\rangle \rbrace \\ =&\min ...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.020230591297149658, -0.02007802203297615, 0.01300646923482418, -0.0037055087741464376, -0.0200627651065588, 0.04689957574009895, 0.03536539152264595, 0.0008934796205721796, 0.005934916902333498, 0.037928543984889984, -0.051873307675123215, -0.0034594919998198748, 0.0017154427478089929, ...
7955f641a12d4bb08d8866315d2bc31dee56ca8f
subsection
29
53
Convergence analysis for RapGrad
In this section, we will first develop the convergence results for Algorithm REF applied to the convex finite-sum subproblem (REF ), and then using them to establish the convergence of RapGrad. Observe that the component functions \psi _i and \varphi in  (REF ) satisfy:\tfrac{\mu }{2}\Vert x-y\Vert ^2\le \psi _i(x)-\ps...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.006579399108886719, -0.0012777383672073483, -0.0004088286077603698, -0.0167364664375782, -0.030787773430347443, 0.029811352491378784, 0.028941728174686432, 0.0026241312734782696, 0.019085979089140892, 0.030955595895648003, -0.031672656536102295, -0.003175275167450309, -0.00868862122297287,...
8d307086c228973acbe4b674b89c9935adfb5696
subsection
30
53
Convergence analysis for RapGrad
Let \hat{\underline{x}}_i^t = (1+\tau _t)^{-1}(\tilde{x}^t+\tau _t\underline{x}_i^{t-1}), for i = 1,\ldots , m,~ t=1,\ldots , s.\mathbb {E}_{i_t} [\psi (\hat{\underline{x}}_i^t)] &= m \psi (\underline{x}_i^t)-(m-1) \psi (\underline{x}_i^{t-1}) ,\\ \mathbb {E}_{i_t} [\nabla \psi (\hat{\underline{x}}_i^t)] &= m\nabla \ps...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1007/s10107-017-1173-0", "end": 1162, "openalex_id": "https://openalex.org/W2964037929", "raw": "Guanghui Lan and Yi Zhou, An optimal randomized incremental gradient method, Mathematical programming, (2017), pp. 1–49.", "source_ref...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.00023537395463790745, 0.00347218313254416, -0.057874806225299835, -0.005860739853233099, 0.009355816058814526, 0.027838513255119324, 0.017246447503566742, 0.03235616907477379, 0.02538127638399601, 0.03424869850277901, -0.03128780424594879, 0.005757718812674284, -0.03690434619784355, 0.0...
102d377e1578ca4e7ac04f229b52234ed0663142
subsection
31
53
Convergence analysis for RapGrad
If the parameters in Algorithm REF satisfy for all t = 1, \ldots , s-1,\alpha _{t+1}\gamma _{t+1}= &~ \gamma _{t}, \\ \gamma _{t+1}[m(1+\tau _{t+1})-1] \le &~ m\gamma _{t}(1+\tau _{t}),\\ \gamma _{t+1}\eta _{t+1}\le &~ \gamma _{t}(1+\eta _{t}),\\ \tfrac{\eta _s \mu }{4}\ge &~ \tfrac{(m-1)^2\hat{L}}{m^2\tau _s},\\ \tfra...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.0003660068614408374, 0.006046862341463566, -0.04681835696101189, -0.02116592600941658, -0.008469422347843647, 0.02243252471089363, 0.011681699194014072, 0.025164103135466576, 0.05078601464629173, 0.043979957699775696, -0.006600045599043369, 0.013406103476881981, -0.021593213081359863, 0....
dd3c9fd58a404afc349fa6ee90935604b9dd6e0a
subsection
32
53
Convergence analysis for RapGrad
If \varphi (x)=\tfrac{\mu }{2}\Vert x-z\Vert ^2, for some z\in X, then, for any s \ge 1, we have\mathbb {E}_s\left[\Vert x^*-x^s\Vert ^2\right]&~\le \textstyle {\alpha ^s(1+2\tfrac{\hat{L}}{\mu })~\mathbb {E}_s\left[\Vert x^*-x^0\Vert ^2+\textstyle {\tfrac{1}{m}\sum _{i=1}^m}\Vert {\underline{x}}^{0}_i-x^0\Vert ^2\righ...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.027728497982025146, -0.008759580552577972, -0.015871018171310425, -0.01924360916018486, -0.03436685726046562, 0.021181704476475716, 0.006825300399214029, 0.046636372804641724, 0.006947385147213936, -0.013780316337943077, -0.01322330441325903, -0.003141774795949459, -0.0036892485804855824,...
bb14e6b9b2d6574dc354f917df457b46a7dfb7d3
subsection
33
53
Convergence analysis for RapGrad
Plugging into (REF ), we obtain the following two relations:\mathbb {E}_s\left[\Vert x^*-x^s\Vert ^2\right]&~\le \alpha ^s\mathbb {E}_s\left[\Vert x^*-x^0\Vert ^2+\textstyle {\sum }_{i=1}^m\tfrac{\hat{L}}{mr}\Vert {\underline{x}}^{0}_i-x^*\Vert ^2\right]\\ &~\le \alpha ^s\mathbb {E}_s\left[\Vert x^*-x^0\Vert ^2+\textst...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.04939063638448715, -0.005899097770452499, -0.0426139235496521, -0.022436413913965225, -0.039011888206005096, 0.03696666285395622, -0.0007350023952312768, 0.04722330719232559, 0.018727537244558334, 0.00598304346203804, -0.0353793241083622, 0.026221605017781258, -0.007097233086824417, 0.0...
b3e41b71413f575e99c065aa08f4762df83c3801
subsection
34
53
Convergence analysis for RapGrad
Actually, as shown below we do not need to solve the subproblem too accurately, and a constant number of iteration of Algorithm REF for each subproblem is enough to guarantee the convergence of Algorithm REF .Let the number of inner iterations s\ge \left\lceil -\log (7M/6)/\log \alpha \right\rceil with M:=\textstyle {...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.017039164900779724, 0.0003096166474279016, -0.015414570458233356, 0.0005591691005975008, -0.026405364274978638, 0.01700865477323532, 0.011128084734082222, -0.031271517276763916, -0.0011974703520536423, 0.012607760727405548, -0.02625282108783722, 0.008397546596825123, -0.015185754746198654...
92834c578d20c434f25999447d1f7c05a18d21be
subsection
35
53
Convergence analysis for RapGrad
According to Theorem REF (with \hat{L} = 2\mu + L), we have, for \ell \ge 1,\mathbb {E}\Vert x_*^{\ell }-\bar{x}^{\ell }\Vert ^2 &\le \textstyle {\alpha ^s(5+\tfrac{2L}{\mu })\mathbb {E}\left[\Vert x^{\ell }_*-\bar{x}^{\ell -1}\Vert ^2+\textstyle {\sum }_{i=1}^m\tfrac{1}{m}\Vert \bar{\underline{x}}^{\ell -1}_i-\bar{x}^...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.04973955079913139, 0.012953643687069416, -0.03313935548067093, -0.04809173569083214, 0.00012182184582343325, 0.049434397369623184, -0.011076967231929302, 0.05721573904156685, -0.005366836208850145, 0.005538483615964651, -0.022169191390275955, 0.041530996561050415, -0.044490959495306015, ...
de971b2dba7283b622932ca1353c403d4ab017a5
subsection
36
53
Convergence analysis for RapGrad
By the optimality condition of the \hat{\ell }-th subproblem (REF ),\nabla \psi ^{\hat{\ell }}(x_*^{\hat{\ell }})+\nabla \varphi ^{\hat{\ell }}(x_*^{\hat{\ell }})\in -N_X(x_*^{\hat{\ell }}).From the definition of \psi ^{\hat{\ell }} and \varphi ^{\hat{\ell }}, we have\nabla f(x_*^{\hat{\ell }})+3\mu (x_*^{\hat{\ell }}-...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.019368596374988556, 0.018391773104667664, -0.03052576445043087, -0.022909585386514664, -0.029121579602360725, 0.03827930614352226, 0.023245370015501976, 0.042858172208070755, 0.025427961722016335, 0.030861547216773033, -0.026984775438904762, 0.023260632529854774, -0.0014013233594596386, ...
61006ed21b739e37d058cf98a990733bfa1b66f6
subsection
37
53
Nonconvex multi-block optimization with linear constraints
In this section, we present a randomized accelerated proximal dual (RapDual) algorithm for solving the nonconvex multi-block optimization problem in (REF ) and show the potential advantages in terms of the total number of block updates.As mentioned in Section 1, we assume the inverse of the last block of the constraint...
{ "cite_spans": [ { "arxiv_id": "", "doi": "10.1137/130936361", "end": 1900, "openalex_id": "https://openalex.org/W1975768153", "raw": "Cong D. Dang and Guanghui Lan, Stochastic block mirror descent methods for nonsmooth and stochastic optimization, SIAM Journal on Optimization, 25 (...
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.010576962493360043, -0.0012419918784871697, 0.006440805736929178, -0.006059241481125355, -0.026602663099765778, 0.06117239221930504, 0.054975785315036774, 0.03379133716225624, 0.00795943196862936, 0.03733225166797638, -0.06434700638055801, -0.0025583887472748756, -0.02706054039299488, 0...
1aea0a59acf5aa3f1ebcb443e8aa096906f87aa0
subsection
38
53
Convergence analysis for RapDual
In this section, we first show the convergence of Algorithm REF for solving the convex multi-block subproblem (REF ) with\psi _i(x)-\psi _i(y)-\langle \nabla \psi _i(y),x-y\rangle \ge \tfrac{\mu }{2}\Vert x-y\Vert ^2, \ \ \forall x, y\in X_i, \quad i=1,\ldots ,m-1, \tfrac{\mu }{2}\Vert x-y\Vert ^2\le \psi _m(x)-\psi _...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ 0.011220754124224186, -0.01322691235691309, -0.02070232853293419, 0.022090621292591095, -0.015233069658279419, 0.02944399043917656, 0.012731092981994152, -0.008482309989631176, 0.039665479212999344, 0.013501519337296486, -0.03270876407623291, 0.005679028574377298, -0.018291888758540154, 0....
d0967fce72ebeeeccee3cc59775de070cfe8c3a4
subsection
39
53
Convergence analysis for RapDual
Let the iterates \emph {\textbf {x}}^t and y^t for t = 1, \ldots , s be generated by Algorithm REF and (\emph {\textbf {x}}^*, y^*) be a saddle point of (REF ). Assume that the parameters in Algorithm REF satisfy for all t = 1, \ldots , s-1\alpha _{t+1} = &~ (m-1) \tilde{\alpha }_{t+1},\\ \gamma _t = &~\gamma _{t+1}\t...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.004168548155575991, -0.010167748667299747, -0.03206693008542061, -0.025537602603435516, -0.00002556477556936443, 0.011151724494993687, 0.00639584194868803, 0.03185335174202919, 0.05360760539770126, 0.01731492206454277, -0.028390368446707726, 0.019023530185222626, -0.00512964092195034, 0...
7e9403e863d88ec1e3186a1fe46e4a4441318e33
subsection
40
53
Convergence analysis for RapDual
Then, for any s \ge 1, we have\mathbb {E}_s\left\lbrace \Vert \emph {\textbf {x}}^s-\emph {\textbf {x}}^*\Vert ^2+\Vert x_m^s-x_m^*\Vert ^2\right\rbrace \le \alpha ^s\mathcal {M}(\Vert \emph {\textbf {x}}^0-\emph {\textbf {x}}^*\Vert ^2 +\Vert x_m^0-x_m^*\Vert ^2),where x_m^*=\operatornamewithlimits{arg\,min}_{x_m\in {...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.008131741546094418, 0.003207689616829157, -0.006987500004470348, -0.014676800929009914, -0.02506651170551777, -0.010069322772324085, -0.005244439002126455, 0.03750060126185417, 0.041986025869846344, 0.016248226165771484, -0.015020073391497135, 0.01585155539214611, 0.005747905466705561, ...
f3e2414c844a69d401a58b319b9b785dd6b99239
subsection
41
53
Convergence analysis for RapDual
Then we have\textstyle {\mathbb {E}_s\left\lbrace \tfrac{(m-1)\gamma _s(\eta _s+\mu )}{2}\Vert {\textbf {x}}^s-{\textbf {x}}^*\Vert ^2+\tfrac{\gamma _s(\tau _s+1)\bar{\mu }}{2}V_{ h}(y^s,y^*)\right\rbrace \le \tfrac{\gamma _1((m-1)\eta _1+(m-2)\mu )}{2}\Vert {\textbf {x}}^0-{\textbf {x}}^*\Vert ^2 +\gamma _1\tau _1V_{...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.039691463112831116, -0.0007627106970176101, -0.008458461612462997, -0.019357597455382347, -0.005811855662614107, 0.01529997680336237, 0.0027019025292247534, 0.030538935214281082, 0.02550504542887211, 0.02431521750986576, -0.03426096588373184, 0.023308439180254936, -0.009732188656926155, ...
f0f4e46e63f70bcf8ab1207f1773fde9bc97e7c2
subsection
42
53
Convergence analysis for RapDual
Also the iterates (\emph {\textbf {x}}^{\ell },x_m^{\ell }) for \ell = 1, \ldots , k be generated by Algorithm REF and \hat{\ell } be randomly selected from [k]. Then\mathbb {E}\left(\Vert {\textbf {x}}^{\ell }_*-\bar{\textbf {x}}^{\ell -1}\Vert ^2+\Vert x_{m^*}^{\ell }-\bar{x}_m^{\ell -1}\Vert ^2\right) & \le \tfrac{1...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.0174453966319561, -0.0017361451173201203, -0.01846800372004509, -0.033578190952539444, -0.012011834420263767, 0.023550521582365036, 0.02187161147594452, 0.028343044221401215, 0.009310316294431686, 0.04700946807861328, -0.024466291069984436, 0.0035772218834608793, -0.024664707481861115, ...
27e09bf0c84a3080c4dac63bb944d2174aed464f
subsection
43
53
Convergence analysis for RapDual
Since ({\textbf {x}}^{\ell }_*,x_{m^*}^{\ell }) is optimal and ({\textbf {x}}_*^{\ell -1}, x_{m^*}^{\ell -1}) is feasible to the \ell -th subproblem, we have\psi ^{\ell }({\textbf {x}}_*^{\ell })+\psi _m^{\ell }(x_{m^*}^{\ell })\le \psi ^{\ell }({\textbf {x}}_*^{\ell -1})+\psi _m^{\ell }(x_{m^*}^{\ell -1}).Plugging in ...
{ "cite_spans": [] }
1805.05411
Accelerated Stochastic Algorithms for Nonconvex Finite-sum and Multi-block Optimization
[ "Guanghui Lan", "Yu Yang" ]
[ "math.OC" ]
2,018
en
Mathematics
[ -0.027972964569926262, 0.012384103611111641, -0.039189621806144714, -0.038060326129198074, -0.02121245115995407, 0.045019228011369705, 0.0013944515958428383, 0.04532444477081299, 0.02086145244538784, 0.05155083164572716, -0.019228551536798477, 0.033726271241903305, -0.005451905075460672, 0...