text
stringlengths
128
2.05k
[MATH] of [MATH] is an induced subgraph of [MATH] formed by a biconnected component [MATH] of [MATH] extended by paths attached to vertices of the biconnected component [MATH] . Each vertex of [MATH] has at most one path attached to it. Each vertex at which a path is attached is a separation vertex in [MATH] and not ad...
Definitions and Basic Terminology We consider graphs [MATH] , where [MATH] is the set of vertices and [MATH] is the set of edges . We distinguish between undirected graphs with edge sets [MATH] and directed graphs with edge sets [MATH] Graph [MATH] is a subgraph of [MATH] if [MATH] and [MATH] . It is an induced subgrap...
A sequence of [MATH] vertices [MATH] [MATH] [MATH] for [MATH] , is an undirected path of length [MATH] , if [MATH] for [MATH] . The vertices [MATH] and [MATH] are the end vertices of undirected path [MATH] . The sequence [MATH] is a directed path of length [MATH] , if [MATH] for [MATH] . Vertex [MATH] is the start vert...
An undirected graph [MATH] is connected if there is a path between every pair of vertices. The distance [MATH] between two vertices [MATH] in a connected undirected graph [MATH] is the smallest integer [MATH] such that there is a path of length [MATH] between [MATH] and [MATH] . A connected component of an undirected g...
[MATH] of [MATH] is an induced biconnected subgraph of [MATH] such that there is no biconnected induced subgraph [MATH] of [MATH] with [MATH] and [MATH]
Definition 2.1 (Resolving set) Let [MATH] be a connected undirected graph. A vertex set [MATH] is a resolving set for [MATH] if for every vertex pair [MATH]
[MATH] , there is a vertex [MATH] such that [MATH] The set [MATH] is a minimum resolving set for [MATH] , if there is no resolving set [MATH] for [MATH] with [MATH] . A connected undirected graph [MATH] has metric dimension
[MATH] if [MATH] is the smallest positive integer such that there is a resolving set for [MATH] of size [MATH] Definition 2.2 Let [MATH] be a connected undirected graph.
1. (leg, root, leaf, hooked leg, ordinary leg) A path [MATH] [MATH] , of [MATH] is a leg , if vertex [MATH] has degree one, the vertices [MATH] have degree [MATH] , and vertex [MATH] has degree [MATH] in [MATH] Vertex [MATH] is called the root of [MATH] . Vertex [MATH] is called the leaf of [MATH] A leg is called a hoo...
2. (bridge) An edge [MATH] is called a bridge if [MATH] is not connected and if [MATH] is not an edge between two vertices of one and the same leg.
3. (extended biconnected component (EBC)) biconnected component [MATH] of [MATH] extended by the subgraphs of [MATH] induced by the vertices of the hooked legs with roots in [MATH] is an extended biconnected component (EBC) of [MATH]
4. (component) Every subgraph induced by the vertices of an ordinary leg, every subgraph induced by the two vertices of a bridge, and every EBC is called a component of [MATH]
5. (amalgamation vertex) Separation vertices of [MATH] that belong to at least two components, i.e. separation vertices without the degree two vertices of the legs and roots of the hooked legs are called amalgamation vertices
Every undirected graph [MATH] can be decomposed into legs, bridges and EBCs. This decomposition is a unique and edge-disjoint partition of [MATH]
Definition 2.3 (EBC-tree) Let [MATH] be a connected undirected graph. 1. The EBC-tree [MATH] for [MATH] is a tree with two types of nodes called c-nodes (nodes for the components of [MATH] ) and a-nodes (nodes for the amalgamation vertices of [MATH] ). [MATH] has a c-node for every component of [MATH] . The vertex set ...
Note that in the EBC-tree all leaves are c-nodes and there is no edge between two a-nodes and no edge between two c-nodes. All ordinary legs are represented by leaves, all bridges are represented by inner c-nodes, and all EBCs are represented by leaves or inner c-nodes.
Definition 2.4 (DEBC-tree) Let [MATH] be a connected undirected graph. 1. For the EBC-tree [MATH] for [MATH] and a node [MATH] let [MATH] be the directed EBC-tree (DEBC-tree) with root [MATH] that is defined as follows: [MATH] contains exactly one directed edge for every undirected edge of [MATH] such that for every no...
2. For a node [MATH] , let [MATH] be the subtree of [MATH] induced by all nodes [MATH] for which there is a directed path from [MATH] to [MATH] in [MATH] . The root of [MATH] is [MATH]
3. For a subtree [MATH] let [MATH] be the set of c-nodes of [MATH] and [MATH] . Then [MATH] is the subgraph of [MATH] induced by the vertices of [MATH]
[MATH] is the subgraph of [MATH] represented by [MATH] . It is not necessary to refer to the a-nodes of [MATH] , because the vertices of [MATH] that are represented by the a-nodes are also represented by the c-nodes since for every a-node [MATH] there is a c-node [MATH] such that \textnu [MATH]
Note that the EBC-tree and the DEBC-tree of [MATH] can be constructed in linear time with the help of any linear time algorithm for finding the biconnected components and bridges of [MATH]
Computing the metric dimension based on a graph decomposition Without loss of generality we will use from now on the following assumptions:
1. [MATH] is a connected undirected, but not biconnected graph. 2. [MATH] is the DEBC-tree for [MATH] with root [MATH] 3. [MATH] is the set of a-nodes of [MATH] and [MATH] is the set of c-nodes of [MATH]
4. Root [MATH] is an a-node. 5. Root [MATH] has at least two children (because [MATH] is not biconnected). First we will describe the general idea of how to compute the metric dimension of [MATH] . The idea is based on dynamic programming.
Property 3.1 For every subtree [MATH] [MATH] , of [MATH] we compute an information [MATH] satisfying the following properties: 1.
For every a-node [MATH] with children [MATH] [MATH] , the information [MATH] can efficiently be computed from [MATH] 2. For every c-node [MATH] with children [MATH] [MATH] , the information [MATH] can efficiently be computed from [MATH] and [MATH]
3. The metric dimension of [MATH] can efficiently be computed from [MATH] These properties allow an efficient bottom-up processing of [MATH] as follows: We start by computing [MATH] for every leaf [MATH] of [MATH] . Since the leafs are c-nodes without children we only need the subgraph [MATH] of [MATH] . For every inne...
Before we define [MATH] we need a few more definitions. Definition 3.2 (Gate Vertex) Let [MATH] be a set of vertices. A vertex [MATH] is an [MATH] -gate of [MATH] , if there is a vertex [MATH] , such that for all [MATH] the equation [MATH] holds. Vertex [MATH] is called an out-vertex for [MATH] -gate [MATH]
Intuitively this definition means that [MATH] is an [MATH] -gate if there is a vertex [MATH] that has a shortest path to any vertex [MATH] that passes [MATH]
Observation 3.3 Let [MATH] and [MATH] be an [MATH] -gate of [MATH] . Then there is an out-vertex [MATH] adjacent to [MATH] Observation 3.4
Let [MATH] [MATH] and [MATH] be an [MATH] -gate of [MATH] . If [MATH] and [MATH] are two out-vertices for [MATH] -gate [MATH] with the same distance to [MATH] , i.e. [MATH] , then both vertices [MATH] and [MATH] have the same distance to all vertices of [MATH] . In this case [MATH] is not a resolving set for [MATH] . C...
Definition 3.5 [MATH] -resolving set, non-gate- [MATH] -resolving set) Let [MATH] 1. [MATH] -resolving set for [MATH] is a resolving set [MATH] for [MATH] with [MATH]
2. minimum [MATH] -resolving set for [MATH] is a resolving set [MATH] for [MATH] with [MATH] such that there is no [MATH] -resolving set [MATH] for [MATH] with [MATH]
3. non-gate- [MATH] -resolving set for [MATH] is a [MATH] -resolving set [MATH] for [MATH] with [MATH] and [MATH] is not an [MATH] -gate in [MATH]
4. minimum non-gate- [MATH] -resolving set for [MATH] is a [MATH] -resolving set [MATH] for [MATH] with [MATH] and [MATH] is not an [MATH] -gate in [MATH] , such that there is no non-gate- [MATH] -resolving set [MATH] for [MATH] with [MATH]
Note that a minimum [MATH] -resolving set is not necessarily a minimum resolving set. It is possible that no minimum resolving set contains [MATH] . Also a minimum non-gate- [MATH] -resolving set is not necessarily a minimum [MATH] -resolving set. It is possible that in every minimum [MATH] -resolving set [MATH] vertex...
Lemma 3.6 Let [MATH] . Let [MATH] a minimum resolving set for [MATH] [MATH] a minimum [MATH] -resolving set for [MATH] , and [MATH] a minimum non-gate- [MATH] -resolving set for [MATH] , then [MATH] and [MATH]
Proof [MATH] : If [MATH] contains vertex [MATH] [MATH] is already a minimum [MATH] -resolving set. If [MATH] the set [MATH] is a [MATH] -resolving set.
[MATH] : If [MATH] is not an [MATH] -gate, [MATH] is already a minimum non-gate- [MATH] -resolving set. If [MATH] is an [MATH] -gate there are out-vertices [MATH] for [MATH] . From Observation 3.4 we know that these out-vertices are on a shortest path between [MATH] and the out-vertex with longest distance to [MATH] , ...
[MATH] Figure shows that these bounds are tight. Lemma 3.7 Let [MATH] be a separation vertex and [MATH] [MATH] , be the vertex sets of the connected components of [MATH] . Let [MATH] be a resolving set for [MATH] . Then there is at most one [MATH] such that [MATH]
Proof Assume there are two sets [MATH] such that [MATH] and [MATH] . Consider a vertex [MATH] adjacent to [MATH] and a vertex [MATH] adjacent to [MATH] . For any [MATH] we have [MATH] . This contradicts the assumption that [MATH] is a resolving set.
[MATH] Lemma 3.8 Let [MATH] be a separation vertex and [MATH] [MATH] , be the vertex sets of the connected components of [MATH] such that if [MATH] in every resolving set [MATH] for [MATH] there is a vertex [MATH] and a vertex [MATH] . Let [MATH] with [MATH] . If [MATH] is a resolving set for [MATH] then [MATH] is reso...
Proof Consider two vertices [MATH] that are separated by [MATH] . We will show, that there is a vertex [MATH] [MATH] , that separates [MATH] and [MATH]
Assume [MATH] and [MATH] are both in the same set [MATH] [MATH] . Then, by assumption and Lemma 3.7 , there is a vertex [MATH] [MATH] . Since every path from [MATH] to [MATH] or [MATH] to [MATH] contains vertex [MATH] and [MATH] separates [MATH] and [MATH] , vertex [MATH] does the same.
Assume [MATH] and [MATH] are in different sets. Without loss of generality let [MATH] [MATH] [MATH] , and [MATH] Case 1: [MATH] If [MATH] then by assumption [MATH] has at least three components. By Lemma 3.7 there is a component [MATH] [MATH] [MATH] , such that [MATH] . Let [MATH] , then we have
[EQUATION] Case 2: [MATH] Then we have for any [MATH] [EQUATION] [MATH] Lemma 3.9 Let [MATH] be a separation vertex and [MATH] [MATH] , be the vertex sets of the connected components of [MATH] such that if [MATH] in every resolving set [MATH] for [MATH] there is a vertex [MATH] and a vertex [MATH] . Let [MATH] [MATH] ,...
[MATH] is a minimum [MATH] -resolving set for [MATH] and there is at most one [MATH] such that [MATH] is an [MATH] -gate in [MATH]
Proof [MATH] : Let [MATH] be a resolving set for [MATH] . We will show that (1) for every [MATH] the set [MATH] is an [MATH] -resolving set for [MATH] and (2) there is at most one index [MATH] such that [MATH] is an [MATH] -gate in [MATH]
(1): Consider two vertices [MATH] . Obviously [MATH] is a resolving set for those pairs of vertices that are separated by a vertex [MATH] Assume that for a pair of vertices [MATH] there is no vertex in [MATH] that separates them, i.e. [MATH] are separated by a vertex [MATH] [MATH] . Since all paths from [MATH] to [MATH...
[EQUATION] Therefore [MATH] implies [MATH] , i.e. [MATH] separates [MATH] and [MATH] (2): Assume there are two indices [MATH] [MATH] , for which [MATH] is an [MATH] -gate in [MATH] and an [MATH] -gate in [MATH] . Then there are out-vertices [MATH] for [MATH] -gate [MATH] in [MATH] and out-vertices [MATH] for [MATH] -ga...
[EQUATION] This contradicts the assumption that [MATH] is a resolving set for [MATH] , see also Figure [MATH] : Let [MATH] be an [MATH] -resolving set for [MATH] and let there be at most one [MATH] such that [MATH] is an [MATH] -gate in [MATH] . We will proof that [MATH] is a resolving set for [MATH]
Let [MATH] . Obviously [MATH] separates all pairs of vertices that are in the same set [MATH] . Consider two vertices [MATH] that are in different set. Without loss of generality let [MATH] and [MATH] [MATH] [MATH] and [MATH] the component in which [MATH] is not an [MATH] -gate for [MATH] . This implies that there is a...
Now we will show the minimality. Assume that [MATH] is a minimal resolving set for [MATH] and [MATH] is not a minimal [MATH] -resolving set, i.e. there is set [MATH] with [MATH] that contains [MATH] and resolves all vertices from [MATH] . Consider the set [MATH] . As shown above [MATH] is a resolving set for [MATH] and...
Assume that [MATH] is a minimal [MATH] -resolving set and there is at most one [MATH] such that [MATH] is an [MATH] -gate in [MATH] , but [MATH] is not minimal, i.e. there is a resolving set [MATH] for [MATH] with [MATH] . This implies that there is at least one [MATH] such that [MATH] is a resolving set for [MATH] and...
[MATH] Lemma 3.10 Let [MATH] be a separation vertex and [MATH] [MATH] , the vertex sets of the connected components of [MATH] such that if [MATH] in every resolving set [MATH] for [MATH] there is a vertex [MATH] and a vertex [MATH] . Let [MATH] . If [MATH] is a minimum resolving set for [MATH] , then [MATH] is a minimu...
Proof Let [MATH] be a minimum resolving set and [MATH] an [MATH] -resolving set. From Lemma 3.8 we know that [MATH] , therefore [MATH] . Assume there is a miniumum [MATH] -resolving set [MATH] with [MATH] . Then, by Lemma 3.8 [MATH] is a resolving set for [MATH] and we have [MATH] . This contradicts the assumption that...
[MATH] Now we define [MATH] [MATH] , as introduced at the beginning of the section. Definition 3.11 1. Let [MATH] be an a-node. We define [MATH] , where [MATH] is the size of a minimum non-gate- \textnu [MATH] -resolving set for [MATH] and [MATH] is the size of a minimum \textnu [MATH] -resolving set for [MATH]
2. Let [MATH] be a c-node with father [MATH] in [MATH] . We define [MATH] , where [MATH] is the size of a minimum non-gate- \textnu [MATH] -resolving set for [MATH] and [MATH] is the size of a minimum \textnu [MATH] -resolving set for [MATH]
To get familiar with this definition we will investigate the smallest possible values for [MATH] and [MATH] . For an arbitrary node [MATH] with father [MATH] we have [MATH] , i.e. the [MATH] -th component of [MATH] is less or equal than the [MATH] -th component of [MATH] , since [MATH] is a subgraph of [MATH] . Therefo...
For an a-node [MATH] that has only leafs as children the graph [MATH] consists of EBCs and paths, that are connected by the separation vertex \textnu [MATH] . Note that if [MATH] has exactly one child [MATH] the graph [MATH] is not an ordinary leg, since this contradicts the decomposition of [MATH] into EBCs, ordinary ...
Observation 3.12 Let [MATH] be a minimum resolving set for [MATH] . For any a-node [MATH] the subgraph [MATH] contains at least one resolving node, i.e. [MATH]
We will now show that this definition satisfys the properties in 3.1 Theorem 3.13 For every a-node [MATH] with children [MATH] [MATH] [MATH] can be computed from [MATH]
Proof [MATH] : If [MATH] has exactly one child [MATH] then [MATH] . Since \textnu [MATH] and [MATH] is the only child of [MATH] , we can follow that [MATH] . Therefore a minimum non-gate- \textnu [MATH] -resolving set for [MATH] is also a minimum non-gate- \textnu [MATH] -resolving set for [MATH] . The same holds for a...
[MATH] : Let [MATH] [MATH] . Then [MATH] with 1. [MATH] 2. [MATH] Let [MATH] be a minimum non-gate- \textnu [MATH] -resolving set for [MATH] and thus [MATH] . Then every [MATH] is also a minimum \textnu [MATH] -resolving set and there is no [MATH] [MATH] , such that [MATH] is a \textnu [MATH] -gate in [MATH] . With the...
In a minimum \textnu [MATH] -resolving set [MATH] for [MATH] there is at most one index [MATH] such that \textnu [MATH] is a [MATH] -gate in [MATH] , otherwise there would be two vertices [MATH] such that for every [MATH] there is a shortest path to [MATH] via \textnu [MATH] , i.e. [MATH] is not a resolving set (see Le...
[MATH] Theorem 3.14 For every c-node [MATH] with father [MATH] and children [MATH] [MATH] [MATH] can be computed from [MATH] and [MATH]
To proof this theorem, we need the following lemma: Lemma 3.15 Let [MATH] be a c-node with father [MATH] and children [MATH] [MATH] . Let [MATH] with \textnu [MATH] . Let [MATH] [MATH] , and [MATH] [MATH] is a \textnu [MATH] -resolving set for [MATH] if and only if
1. [MATH] is a resolving set for [MATH] and 2. [MATH] is a \textnu [MATH] -resolving set for [MATH] and 3. For every [MATH] vertex \textnu [MATH] is not an [MATH] -gate in [MATH] or not an [MATH] -gate in [MATH]
Proof [MATH] ”: Let [MATH] be a \textnu [MATH] -resolving set for [MATH] 1. We show that [MATH] is a resolving set for [MATH] . Let [MATH] . Pair [MATH] is either resolved by a vertex [MATH] , or by a vertex [MATH] , or by a vertex [MATH] [MATH] [MATH] . Since every path from a vertex of [MATH] to [MATH] or to [MATH] c...
2. We show that [MATH] is a \textnu [MATH] -resolving set for [MATH] . Since \textnu [MATH] by definition, we just have to show that [MATH] is a resolving set. Let [MATH] . Pair [MATH] is either resolved by a vertex [MATH] or by a vertex [MATH] [MATH] . Since every path from a vertex in [MATH] to [MATH] contains vertex...
3. We show that for every [MATH] vertex \textnu [MATH] is not an [MATH] -gate in [MATH] or not an [MATH] -gate in [MATH] . Assume there is an index [MATH] such that vertex \textnu [MATH] is an [MATH] -gate in [MATH] and an [MATH] -gate in [MATH] . From Observation 3.3 we know that there is an out-vertex [MATH] with res...
[MATH] ”: Let 1. [MATH] be a resolving set for [MATH] 2. [MATH] be a \textnu [MATH] -resolving set for [MATH] 3. for every [MATH] vertex \textnu [MATH] be not an [MATH] -gate in [MATH] or not an [MATH] -gate in [MATH]
We show that [MATH] is a \textnu [MATH] -resolving set. [MATH] is a \textnu [MATH] -resolving set, if \textnu [MATH] and if for every pair [MATH] there is a vertex in [MATH] , that resolves [MATH] and [MATH] . Obviously \textnu [MATH] , because \textnu [MATH] , so we just have to show that [MATH] is a resolving set. We...
(a) [MATH] [MATH] (b) [MATH] [MATH] [MATH] [MATH] (c) [MATH] (d) [MATH] and [MATH] [MATH] We will have a closer look at all these cases.
(a) Let [MATH] . Pair [MATH] is either resolved by a vertex [MATH] , or by vertex \textnu [MATH] . Since [MATH] and every pair that is resolved by \textnu [MATH] is also resolved by a vertex in [MATH] \textnu [MATH] ), [MATH] resolves all pairs [MATH] , see Figure 7(a)
(b) Let [MATH] and [MATH] . Pair [MATH] is either resolved by [MATH] or by [MATH] , because it is not possible that [MATH] and [MATH] have the same distance to both, [MATH] and [MATH] , what can be seen as follows:
Let [MATH] [MATH] [MATH] [MATH] [MATH] [MATH] and [MATH] , (see Figure 7(b) ). Since [MATH] we can follow [MATH] . Assume that neither [MATH] nor [MATH] separate [MATH] , i.e. [MATH] and [MATH] . From [MATH] and [MATH] we get [MATH] and [MATH] . Finally we have [MATH] what implies [MATH] and thus [MATH] . This contradi...
(c) Let [MATH] . Pair [MATH] is either resolved by a vertex [MATH] or by a vertex in [MATH] . Since [MATH] for every [MATH] (see Observation 3.12 ) and every pair that is resolved by \textnu [MATH] is also resolved by a vertex in [MATH] [MATH] resolves all pairs [MATH] , see Figure 7(c)
(d) Let [MATH] and [MATH] . Since for every [MATH] vertex \textnu [MATH] is not an [MATH] -gate in [MATH] or not an [MATH] -gate in [MATH] , pair [MATH] is either resolved by a vertex [MATH] (see Observation 3.12 ) or by a vertex [MATH] \textnu [MATH] ), what can be seen as follows:
Assume pair [MATH] cannot be resolves by a vertex [MATH] and cannot be resolved by a vertex [MATH] . Without loss of generality let \textnu [MATH] be no [MATH] -gate in [MATH] . Then there is a vertex [MATH] such that there is no shortest path from [MATH] to [MATH] via vertex \textnu [MATH] Therefore [MATH] . Since [MA...
Vertex [MATH] does not resolve pair [MATH] either, that means [MATH] . It follows [MATH] , what contradicts the asumption. [MATH]
Proof of Theorem 3.14 Graph [MATH] is composed by the graph [MATH] and the graphs [MATH] [MATH] . We compute [MATH] by computing a minimum-non-gate- \textnu [MATH] -resolving set [MATH] for [MATH] with [MATH] and a minimum \textnu [MATH] -resolving set [MATH] for [MATH] with [MATH] with the help of Lemma 3.15
Let [MATH] be a minimum-non-gate- \textnu [MATH] -resolving set for [MATH] and [MATH] be a minimum- \textnu [MATH] -resolving set for [MATH] [MATH] . To compute sets [MATH] and [MATH] and thus [MATH] and [MATH] we can do the following:
For every subset [MATH] that contains vertices \textnu [MATH] [MATH] and resolves all pairs [MATH] we determine a resolving set [MATH] for [MATH]
[MATH] contains the vertices in [MATH] and for every [MATH] either the vertices in [MATH] or in [MATH] . If vertex \textnu [MATH] is a [MATH] -gate in [MATH] then [MATH] contains the vertices in [MATH] else the vertices in [MATH] [MATH] is a \textnu [MATH] -resolving set for [MATH] (Lemma 3.15 ) and by Lemma 3.8 we get...
Then we have [MATH] with [MATH] and [MATH] with [MATH] [MATH] Theorem 3.16 The metric dimension of [MATH] can efficiently be computed from [MATH]
Proof Let [MATH] be a minimum- \textnu [MATH] -resolving set for [MATH] . Since the conditions of Lemma 3.8 are given, we get that [MATH] is a resolving set for [MATH] and with Lemma 3.6 we get that [MATH] is a minimum resolving set for [MATH] . Thus, if [MATH] , then [MATH] is the metric dimension of [MATH]
[MATH] Algorithm and Time Complexity Let [MATH] be a connected undirected graph with [MATH] and [MATH] . To compute a resolving set for [MATH] we first compute the DEBC-tree [MATH] for [MATH] . This can be done in [MATH] with the help of any linear-time-algorithm for finding the biconnected components and bridges of [M...
Definition 4.1 An undirected graph [MATH] is (minimum) [MATH] -EBC-bounded for some positive integer [MATH] , if there is a (minimum) resolving set [MATH] for [MATH] such that every EBC of [MATH] contains at most [MATH] vertices of [MATH] [MATH] is called a (minimum) [MATH] -EBC-bounded-resolving set for [MATH]
Let [MATH] and [MATH] be the class of graphs that are [MATH] -EBC-bounded and minimum- [MATH] -EBC-bounded, respectively. A set of graphs [MATH] is (minimum) EBC-bounded , if for every graph [MATH] there is a [MATH] such that [MATH] is (minimum) [MATH] -EBC-bounded.
Corollary 4.2 The following problems can be solved in polynomial time for any fixed positive integer [MATH] 1. Given an undirected graph [MATH] . Is [MATH]
2. Given a set [MATH] of EBC-bounded graphs. Find the smallest integer [MATH] such that [MATH] 3. Given an undirected graph [MATH] . Compute a minimum [MATH] -EBC-bounded-resolving set for [MATH]
4. Given an undirected graph [MATH] . Compute a minimum resolving set for [MATH] and thus the metric dimension of [MATH] To solve these problems we use our algorithm with slight modifications.
Instead of checking every subset [MATH] if it is resolving, we do the following: For the problems 1., 3. and 4. we only test those subsets with at most [MATH] vertices. For the problem 2. we run our algorithm for [MATH] and increase [MATH] successively by one until we get a resolving set. By doing these modifications t...
Obviously it holds that [MATH] . Vice versa it holds that for all [MATH] there is a graph [MATH] such that [MATH] , see Figure . Moreover, the complexity of the following problems remain open:
1. Given an undirected graph [MATH] a fixed positive integer [MATH] . Is [MATH] 2. Given an undirected graph [MATH] . Find the smallest integer [MATH] such that [MATH]
The following problem, however, still remains NP-complete. The proof can be found in the full version of this paper. [MATH] -bounded BC Metric Dimension
Given: An undirected graph [MATH] and a positive integer [MATH] such that there is a minimum resolving set [MATH] for [MATH] that contains at most [MATH] vertices from each biconnected component of [MATH]
Question: Is the metric dimension of [MATH] at most [MATH] Theorem 4.3 [MATH] -bounded BC Metric Dimension is NP-complete for all positive integers [MATH]
Proof We use a slight modification of the NP-completeness proof of Metric Dimension from Khuller et al. in , where they reduce from 3-SAT. Let [MATH] be a 3-SAT instance with [MATH] variables and [MATH] clauses. For each variable [MATH] they construct the gadget in Figure 9(a) and for each clause [MATH] the gadget in F...