This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Graphs with the edge metric dimension smaller than the metric dimension footnotetext: [email protected], [email protected], [email protected], [email protected], and
      [email protected]

Martin Knor1, Snježana Majstorović2, Aoden Teo Masa Toshi3,
Riste Škrekovski4,5 and Ismael G. Yero6
1 Slovak University of Technology in Bratislava, Bratislava, Slovakia
2 University of Osijek, Department of Mathematics, Croatia
3 Independent researcher, Singapore
4 University of Ljubljana, FMF, 1000 Ljubljana, Slovenia
5 Faculty of Information Studies, 8000 Novo Mesto, Slovenia
6 Universidad de Cádiz, Departamento de Matemáticas, Algeciras, Spain
Abstract

Given a connected graph GG, the metric (resp. edge metric) dimension of GG is the cardinality of the smallest ordered set of vertices that uniquely identifies every pair of distinct vertices (resp. edges) of GG by means of distance vectors to such a set. In this work, we settle three open problems on (edge) metric dimension of graphs. Specifically, we show that for every r,t2r,t\geq 2 with rtr\neq t, there is n0n_{0}, such that for every nn0n\geq n_{0} there exists a graph GG of order nn with metric dimension rr and edge metric dimension tt, which among other consequences, shows the existence of infinitely many graph whose edge metric dimension is strictly smaller than its metric dimension. In addition, we also prove that it is not possible to bound the edge metric dimension of a graph GG by some constant factor of the metric dimension of GG.

Keywords: Edge metric dimension; metric dimension; unicyclic graphs.

AMS Subject Classification numbers: 05C12; 05C76

1 Introduction

Metric dimension is nowadays a well studied topic in graph theory and combinatorics, as well as in some computer science applications, and the theory involving it is indeed full of interesting results and open questions. One recent issue that has attracted the attention of several researchers concerns a variant of the standard metric dimension, in which it is required to uniquely recognize the edges of a graph, instead of its vertices, and by using vertices as the recognizing elements. This variant was introduced in [6], and since its appearance, a significant number of works have been published. In this sense, we mention the most recent ones [3, 4, 9, 12, 13, 14]. Specifically, the edge metric dimension is studied in several situations as follows: [3] is dedicated to study several generalized Petersen graphs; in [4], a number of results about pattern avoidance in graphs with bounded edge metric dimension are given; [9] centers the attention on some product graphs (corona, join and lexicographic); [12] studies some convex polytopes and related graphs; in [13], a characterization of graphs with the largest possible edge metric dimension (order minus one) is given; and finally, in [14] the Cartesian product of any graph with a path is studied, as well as, it is proved to be not possible to bound the metric dimension of a graph GG by some constant factor of the edge metric dimension of GG. We should also remark some results from the seminal article [6]. There was proved for instance that computing the edge metric dimension of graphs is NP-hard, that can be approximated within a constant factor, and that becomes polynomial for the case of trees. Further, some bounds and closed formulas for several classes of graphs including trees, grid graphs and wheels (among others), were also deduced in [6].

We recall that the parameter edge metric dimension (from [6]) studied here is not the same as that one defined in [8], where the authors studied the metric dimension of the line graph of a graph (namely edges uniquely recognizing edges), and called such parameter as edge metric dimension, although it was further renamed as the edge version of metric dimension in [7].

In the next we recall the necessary terminology and notation. We consider only simple and connected graphs. Let GG be a graph and let u,vu,v be its vertices. By dG(u,v)d_{G}(u,v) (or by d(u,v)d(u,v) when no confusion is likely) we denote the distance from uu to vv in GG. Let zz be a vertex of GG. We say that zz identifies (resolves or determines) a pair of vertices u,vV(G)u,v\in V(G), if dG(u,z)dG(v,z)d_{G}(u,z)\neq d_{G}(v,z), An ordered set of vertices SS is a metric generator for GG if every two vertices u,vV(G)u,v\in V(G) are identified by a vertex of SS. The metric dimension of GG is then the cardinality of the smallest metric generator for GG. Such cardinality is denoted by dim(G)\dim(G) and a metric generator of cardinality dim(G)\dim(G) is known as a metric basis. It is necessary to remark that the concepts above were first defined in [1] for a more general setting of metric spaces. The concepts were again independently rediscovered for the case of graphs in [5] and [11], where metric generators were called resolving sets and locating sets, respectively. Also, in [11], the metric dimension was called locating number. The terminology of metric generators was first used in [10].

Let GG be a graph and let SS be an ordered set of vertices of GG. For every vV(G)v\in V(G) we can consider the vector r(v|S)r(v|S) of distances from vv to the vertices in SS. If SS is a metric generator, then all such vectors are pairwise different. The vector r(v|S)r(v|S) is known as the metric representation of vv with respect to SS.

The concept of edge metric dimension was first described in [6], as a way to uniquely recognize the edges of a given graph GG. A vertex zV(G)z\in V(G) distinguishes two edges e,fE(G)e,f\in E(G) if dG(e,z)dG(f,z)d_{G}(e,z)\neq d_{G}(f,z), where dG(e,z)=dG(uv,z)=min{dG(u,z),dG(v,z)}d_{G}(e,z)=d_{G}(uv,z)=\min\{d_{G}(u,z),d_{G}(v,z)\}. A set of vertices SV(G)S\subset V(G) is an edge metric generator for GG, if any two edges of GG are distinguished by a vertex of SS. The edge metric dimension of GG is the cardinality of the smallest edge metric generator for GG, and is denoted by edim(G){\rm edim}(G). An edge metric generator of cardinality edim(G){\rm edim}(G) is known as an edge metric basis. The edge metric representation is defined analogously as in the case of the metric dimension.

2 Edge metric dimension versus metric dimension

One would expect that dim(G)\dim(G) and edim(G){\rm edim}(G) are related. The search for a relationship between these two invariants (in a shape of a bound for instance) was of interest in the seminal article [6], as well as in the subsequent works on the topic (see also for instance [14]). In this sense, in [6], several families of graphs for which dim(G)<edim(G)\dim(G)<{\rm edim}(G), or dim(G)=edim(G)\dim(G)={\rm edim}(G), or dim(G)>edim(G)\dim(G)>{\rm edim}(G) were presented. For the last case, only one construction was given, namely the Cartesian product of two cycles C4rC4tC_{4r}\Box C_{4t}. It was shown in [6] that edim(C4rC4t)=3<4=dim(C4rC4t){\rm edim}(C_{4r}\Box C_{4t})=3<4=\dim(C_{4r}\Box C_{4t}). In consequence, it was claimed in [6] that the metric dimension and the edge metric dimension of graphs seemed to be not comparable in general. This example above and the other results from [6] allowed the authors of that article to point out the following questions.

  • (i)

    For which integers r,t,n1r,t,n\geq 1, with r,tn1r,t\leq n-1, can be constructed a graph GG of order nn with dim(G)=r\dim(G)=r and edim(G)=t{\rm edim}(G)=t?

  • (ii)

    Is it possible that dim(G)\dim(G) or edim(G){\rm edim}(G) would be bounded from above by some constant factor of edim(G){\rm edim}(G) or dim(G)\dim(G), respectively?

  • (iii)

    Can you construct some other families of graphs for which dim(G)>edim(G)\dim(G)>{\rm edim}(G)?

Note that the question (i) is precisely the realization of graphs GG with prescribed values on its order, metric dimension and edge metric dimension, while the question (ii) is equivalent to ask whether the ratios edim(G)dim(G)\frac{{\rm edim}(G)}{\dim(G)} and dim(G)edim(G)\frac{\dim(G)}{{\rm edim}(G)} are bounded from above. The third question can be settled as a consequence of the other two. Realization results concerning the case in which dim(G)edim(G)\dim(G)\leq{\rm edim}(G) were already studied in [6], although not completed. With respect to the ratios, it was proved in [14] that edim(G)dim(G)\frac{{\rm edim}(G)}{\dim(G)} is not bounded from above. The other possibility has never been studied till now.

In this work we deal with these three problems mentioned above. That is, our results complete the unboundedness results given in [14], while studying the ratio dim(G)edim(G)\frac{\dim(G)}{{\rm edim}(G)}, and thus, the problem in (ii) is now completely settled. We also give positive answer to (iii), and moreover, we present an almost complete answer to (i), since we show that for every r,t2r,t\geq 2 with rtr\neq t, there is n0n_{0}, such that for every nn0n\geq n_{0} there exists an outerplanar graph GG (a cactus graph indeed), of order nn with dim(G)=r\dim(G)=r and edim(G)=t{\rm edim}(G)=t. This result is in a sense best possible, because if edim(G)=1{\rm edim}(G)=1, then GG is a path of length at least 22, and consequently dim(G)=1\dim(G)=1 as well. We remark that a similar result for 2rt2r2\leq r\leq t\leq 2r was proved in [6], where n0n_{0} was shown to be at most 2r+22r+2. So, our result complements the former one and a weaker version of this former result (with a weaker bound for n0n_{0}) can be proved also using a variant of our construction.

As a consequence of our results it is clear the existence of infinite families of graphs GG for which the differences edim(G)dim(G){\rm edim}(G)-\dim(G) and dim(G)edim(G)\dim(G)-{\rm edim}(G) are arbitrarily large. Proving that the difference edim(G)dim(G){\rm edim}(G)-\dim(G) is arbitrarily large was already presented in [6]. However, the other difference dim(G)edim(G)\dim(G)-{\rm edim}(G) was only proved to be at most 1 in [6], and there was no more knowledge on this issue. Clearly, the unboundedness of the ratio dim(G)edim(G)\frac{\dim(G)}{{\rm edim}(G)} gives as a consequence that dim(G)edim(G)\dim(G)-{\rm edim}(G) can as large as possible.

While graphs for which dim(G)<edim(G)\dim(G)<{\rm edim}(G) are very common, and they are present in several investigations already published, the opposed version dim(G)<edim(G)\dim(G)<{\rm edim}(G) seemed to be more elusive till now. We have first observed that K2K_{2} is the unique connected simple graph whose edge metric dimension is 0. Since dim(K2)=1\dim(K_{2})=1, K2K_{2} is the smallest graph which has the edge metric dimension smaller than the metric dimension. For non-trivial examples one needs to consider graphs of order at least 10. By exhaustive computer search we found that the smallest possible graphs GG (different from K2K_{2}) satisfying the inequality edim(G)<dim(G){\rm edim}(G)<\dim(G), are the five graphs on 1010 vertices depicted in Figure 4. Moreover, we have found 61 such graphs of order 11.

Refer to caption
Figure 1: The smallest graphs GG for which edim(G)<dim(G){\rm edim}(G)<\dim(G), different from K2K_{2}. The squared vertices form a metric basis and the circled bolded vertices form an edge metric basis.

The main contributions of our work are as follows.

Theorem 1.

Let k1,k22k_{1},k_{2}\geq 2 and k1k2k_{1}\neq k_{2}. Then there is an integer n0n_{0} such that for every nn0n\geq n_{0} there exists a graph on nn vertices with dim(G)=k1\dim(G)=k_{1} and edim(G)=k2{\rm edim}(G)=k_{2}.

Theorem 2.

The ratio dim(G)edim(G)\frac{\dim(G)}{{\rm edim}(G)} is not bounded from above.

Proof.

By Theorem 1, for arbitrarily large NN, it is always possible to find a graph GG such that dim(G)=Nk\dim(G)=Nk and edim(G)=k{\rm edim}(G)=k. As such, dim(G)edimG\frac{\dim(G)}{{\rm edim}{G}} can be made arbitrarily large. ∎

Notice that by using a similar argument as the one in the proof above, although it is already known from [14], we can also prove that the ratio edim(G)dim(G)\frac{{\rm edim}(G)}{\dim(G)} is not bounded from above.

3 Proof of Theorem 1

In order prove the main result of this work, we shall construct infinite families of graphs GG for which edim(G)<dim(G){\rm edim}(G)<\dim(G) as well as other ones where dim(G)<edim(G)\dim(G)<{\rm edim}(G). To this end, we need first some preliminary results. We first describe two graphs that will be used in this purpose. Take a cycle CC on n1n_{1} vertices, where n15n_{1}\geq 5. We denote the vertices of CC consecutively by a1,a2,,an1a_{1},a_{2},\dots,a_{n_{1}}. Further, take a path PP on n2n_{2} vertices denoted consecutively by b1,b2,,bn2b_{1},b_{2},\dots,b_{n_{2}}, where n21n_{2}\geq 1, and join PP to CC by the edge a2b1a_{2}b_{1}. Then take vertices cc and ii and connect them by edges to an1a_{n_{1}} and a1a_{1}, respectively. Finally, take n3n_{3} vertices j1,j2,,jn3j_{1},j_{2},\dots,j_{n_{3}}, where n32n_{3}\geq 2, and join them by edges to the vertex ii. We denote the resulting graph by Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. A fairly representative example of a graph as described above in drawn in Figure 2.

Refer to caption
Figure 2: The graph G7,3,4G_{7,3,4}

In connection with the graphs Gn1,n2,n3G_{n_{1},n_{2},n_{3}}, we shall need a graph denoted by Gn1,n2G_{n_{1},n_{2}} that is obtained from Gn1,n2,n3G_{n_{1},n_{2},n_{3}} by removing the vertices of the set {i,j1,j2,,jn3}\{i,j_{1},j_{2},\dots,j_{n_{3}}\} and the edges incident with them. Thus, Gn1,n2G_{n_{1},n_{2}} is a proper subgraph of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}.

3.1 Preliminaries of the proof

We now prove several auxiliary results about the graph Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. We remark that we shall be using the graphs Gn1,n2G_{n_{1},n_{2}} in most of the described situations.

Observation 3.

Let n15n_{1}\geq 5, n21n_{2}\geq 1 and n32n_{3}\geq 2. Then dim(Gn1,n2,n3)n3\dim(G_{n_{1},n_{2},n_{3}})\geq n_{3} and edim(Gn1,n2,n3)n3{\rm edim}(G_{n_{1},n_{2},n_{3}})\geq n_{3}.

Proof.

Let SS be a metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. If SS does not contain a vertex of the subgraph Gn1,n2G_{n_{1},n_{2}}, then r(a2|S)=r(an1|S)r(a_{2}|S)=r(a_{n_{1}}|S). Thus, SS contains at least one vertex of Gn1,n2G_{n_{1},n_{2}}. Moreover, if SS does not contain two pendant vertices jrj_{r} and jtj_{t} attached to ii, then r(jr|S)=r(jt|S)r(j_{r}|S)=r(j_{t}|S). Thus, SS contains at least n31n_{3}-1 pendant vertices attached to ii, and so, in conclusion we get dim(Gn1,n2,n3)n3\dim(G_{n_{1},n_{2},n_{3}})\geq n_{3}.

Now let TT be an edge metric basis. Here the situation is analogous since r(a1a2|T)=r(a1an1|,T)r(a_{1}a_{2}|T)=r(a_{1}a_{n_{1}}|,T) if TT does not contain a vertex of Gn1,n2G_{n_{1},n_{2}}, and r(jri|T)=r(jti|T)r(j_{r}i|T)=r(j_{t}i|T) if jrj_{r} and jtj_{t} are pendant vertices attached to ii and not in TT. ∎

Lemma 4.

Let n15n_{1}\geq 5, n21n_{2}\geq 1 and n32n_{3}\geq 2. Then dim(Gn1,n2,n3)n3+1\dim(G_{n_{1},n_{2},n_{3}})\leq n_{3}+1 and edim(Gn1,n2,n3)n3+1{\rm edim}(G_{n_{1},n_{2},n_{3}})\leq n_{3}+1.

Proof.

Let S={j1,j2,,jn31,aα,aβ}S=\{j_{1},j_{2},\dots,j_{n_{3}-1},a_{\alpha},a_{\beta}\}, where α=n1+12\alpha=\lfloor\frac{n_{1}+1}{2}\rfloor and β=n1+32\beta=\lceil\frac{n_{1}+3}{2}\rceil. Observe that d(a1,aα)=d(a1,aβ)d(a_{1},a_{\alpha})=d(a_{1},a_{\beta}) since d(a1,aα)=n1+121d(a_{1},a_{\alpha})=\lfloor\frac{n_{1}+1}{2}\rfloor-1 and d(a1,aβ)=n1+1n1+32d(a_{1},a_{\beta})=n_{1}+1-\lceil\frac{n_{1}+3}{2}\rceil. Moreover, d(aα,aβ)=1d(a_{\alpha},a_{\beta})=1 if n1n_{1} is odd and d(aα,aβ)=2d(a_{\alpha},a_{\beta})=2 otherwise. Obviously, SS contains n3+1n_{3}+1 vertices. And since n15n_{1}\geq 5, we have β<n1\beta<n_{1}. Denote γ=d(aα,a1)=n112\gamma=d(a_{\alpha},a_{1})=\lfloor\frac{n_{1}-1}{2}\rfloor.

First we show that SS is a metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. Since n32n_{3}\geq 2, there is a vertex of SS outside the subgraph Gn1,n2G_{n_{1},n_{2}}. This vertex distinguishes those vertices of Gn1,n2G_{n_{1},n_{2}} which are at different distances from a1a_{1}, but not those which are at the same distance from a1a_{1}. According to the distance from a1a_{1}, the vertices of Gn1,n2G_{n_{1},n_{2}} are partitioned into nontrivial sets {a2,an1}\{a_{2},a_{n_{1}}\}, {a3,an11,b1,c}\{a_{3},a_{n_{1}-1},b_{1},c\}, {a4,an12,b2}\{a_{4},a_{n_{1}-2},b_{2}\}, etc. Some of these sets are probably smaller since they do not need to contain vertices of both CC and PP, and n1n_{1} could be even.

Let δ=n12\delta=\lfloor\frac{n_{1}}{2}\rfloor. For every vV(Gn1,n2)v\in V(G_{n_{1},n_{2}}), denote r~(v)=(d(v,aα),d(v,aβ))\tilde{r}(v)=(d(v,a_{\alpha}),d(v,a_{\beta})). Then the vertices in the first set of the partition have r~(a2)=(γ1,δ)\tilde{r}(a_{2})=(\gamma-1,\delta) and r~(an1)=(δ,γ1)\tilde{r}(a_{n_{1}})=(\delta,\gamma-1). Since γδ\gamma\leq\delta, these pairs are different. Further, the pairs r~\tilde{r} for vertices in the second set of the partition are (γ2,δ1)(\gamma-2,\delta-1), (δ1,γ2)(\delta-1,\gamma-2), (γ,δ+1)(\gamma,\delta+1), (δ+1,γ)(\delta+1,\gamma), and they are pairwise distinct too. Since the distances to vertices of CC decrease while the distances to vertices of PP increase, it is obvious that the pair aα,aβa_{\alpha},a_{\beta} distinguishes the vertices in the sets of the partition. Hence, SS identifies the vertices of Gn1,n2G_{n_{1},n_{2}}.

Since a pendant vertex in SS is identified trivially, it remains to check the vertices ii and jn3j_{n_{3}}. Obviously, jn3j_{n_{3}} is the only vertex at distance 22 from the pendant vertices in SS and at distance γ+2\gamma+2 from aαa_{\alpha}. On the other hand, ii is the only vertex at distance 11 from the pendant vertices in SS. Thus, SS is a metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}.

Now we show that SS is an edge metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. We proceed analogously as in the case for the metric generator. According to the distance from a1a_{1}, the vertices of SS outside Gn1,n2G_{n_{1},n_{2}} partition the edges of GG into sets {a1a2,a1an1}\{a_{1}a_{2},a_{1}a_{n_{1}}\}, {a2a3,an1an11,a2b1,an1c}\{a_{2}a_{3},a_{n_{1}}a_{n_{1}-1},a_{2}b_{1},a_{n_{1}}c\}, {a3a4,an11an12,b1b2}\{a_{3}a_{4},a_{n_{1}-1}a_{n_{1}-2},b_{1}b_{2}\}, etc. Again, some of these sets are probably smaller since they do not need to contain the edges of both CC and PP, and n1n_{1} may be odd. For every eE(Gn1,n2,n3)e\in E(G_{n_{1},n_{2},n_{3}}), denote r~(v)=(d(e,aα),d(e,aβ))\tilde{r}(v)=(d(e,a_{\alpha}),d(e,a_{\beta})). Then the pairs r~\tilde{r} for edges in the first set of the partition are (γ1,γ)(\gamma-1,\gamma) and (γ,γ1)(\gamma,\gamma-1). Further, the pairs r~\tilde{r} for edges in the second set of the partition are (γ2,δ1)(\gamma-2,\delta-1), (δ1,γ2)(\delta-1,\gamma-2), (γ1,δ)(\gamma-1,\delta), (δ,γ1)(\delta,\gamma-1) and they are pairwise distinct too. Since the distances to edges of CC decrease while the distances to edges of PP increase, it is obvious that the pair aα,aβa_{\alpha},a_{\beta} distinguishes the edges in the sets of the partition. Hence, SS identifies the edges of Gn1,n2G_{n_{1},n_{2}}.

Since a pendant edge ijtij_{t} is identified trivially if jtSj_{t}\in S, it remains to check the edges ia0ia_{0} and ijn3ij_{n_{3}}. Obviously, ijn3ij_{n_{3}} is the only edge at distance 11 from the pendant vertices in SS and at distance γ+1\gamma+1 from aαa_{\alpha}. On the other hand, ia1ia_{1} is the only edge at distance 11 from the pendant vertices in SS and at distance γ\gamma from aαa_{\alpha}. Therefore, SS is a metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} and the proof is completed. ∎

The next two propositions show that dim(Gn1,n2,n3)\dim(G_{n_{1},n_{2},n_{3}}) and edim(Gn1,n2,n3){\rm edim}(G_{n_{1},n_{2},n_{3}}) depend on the parity of n1n_{1}.

Lemma 5.

Let n15n_{1}\geq 5, n21n_{2}\geq 1 and n32n_{3}\geq 2. Then dim(Gn1,n2,n3)=n3\dim(G_{n_{1},n_{2},n_{3}})=n_{3} if n1n_{1} is odd, and dim(Gn1,n2,n3)=n3+1\dim(G_{n_{1},n_{2},n_{3}})=n_{3}+1 if n1n_{1} is even.

Proof.

First assume that n1n_{1} is odd. Analogously as in the proof of Lemma 4, let α=n1+12\alpha=\lfloor\frac{n_{1}+1}{2}\rfloor, γ=n112\gamma=\lfloor\frac{n_{1}-1}{2}\rfloor and δ=n12\delta=\lfloor\frac{n_{1}}{2}\rfloor. Since n1n_{1} is odd, γ=δ\gamma=\delta. As shown in Observation 3, every metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} contains a vertex outside Gn1,n2G_{n_{1},n_{2}}. By using the distances, this vertex partitions V(Gn1,n2)V(G_{n_{1},n_{2}}) into nontrivial sets P1={a2,an1}P_{1}=\{a_{2},a_{n_{1}}\}, P2={a3,an11,b1,c}P_{2}=\{a_{3},a_{n_{1}-1},b_{1},c\}, P3{a4,an12,b2}P_{3}\subseteq\{a_{4},a_{n_{1}-2},b_{2}\}, etc. For every vV(Gn1,n2)v\in V(G_{n_{1},n_{2}}), we denote r~(v)=d(aα,v)\tilde{r}(v)=d(a_{\alpha},v). Then the values of r~\tilde{r} for the vertices in P1P_{1} are γ1\gamma-1 and δ\delta, and they are different. The values of r~\tilde{r} for vertices in P2P_{2} are γ2\gamma-2, δ1\delta-1, γ\gamma and δ+1\delta+1, and they are different too. The set P3P_{3} contains vertices with values of r~\tilde{r} being γ3\gamma-3, δ2\delta-2 and γ+1\gamma+1, etc. Hence, to identify the vertices of Gn1,n2G_{n_{1},n_{2}} it suffices (and is necessary by the proof of Observation 3) that a metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} will contain only one vertex of Gn1,n2G_{n_{1},n_{2}}, namely aαa_{\alpha}. Hence, {j1,j2,,jn31,aα}\{j_{1},j-2,\dots,j_{n_{3}-1},a_{\alpha}\} is a metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}, see the proof of Lemma 4. By Observation 3, we then get dim(Gn1,n2,n3)=n3\dim(G_{n_{1},n_{2},n_{3}})=n_{3}.

We now assume n1n_{1} is even and proceed analogously as above. Vertices outside Gn1,n2G_{n_{1},n_{2}} partition V(Gn1,n2,n3)V(G_{n_{1},n_{2},n_{3}}) into nontrivial sets P1={a2,an1}P_{1}=\{a_{2},a_{n_{1}}\}, P2={a3,an11,b1,c}P_{2}=\{a_{3},a_{n_{1}-1},b_{1},c\}, P3{a4,an12,b2}P_{3}\subseteq\{a_{4},a_{n_{1}-2},b_{2}\}, etc. We denote this partition by 𝒫\mathcal{P}. (Though it is not obvious, this partition is slightly different from the partition for the case when n1n_{1} is odd.) We show that there is not a unique vertex in Gn1,n2G_{n_{1},n_{2}} which distinguishes all the vertices inside the sets of 𝒫\mathcal{P}. Let vV(Gn1,n2,n3)v\in V(G_{n_{1},n_{2},n_{3}}). By symmetry, it suffices to distinguish four cases:

Case 1: v{a1,an1+22}v\in\{a_{1},a_{\frac{n_{1}+2}{2}}\}. Then d(v,a2)=d(v,an1)d(v,a_{2})=d(v,a_{n_{1}}) and a2,an1P1a_{2},a_{n_{1}}\in P_{1}.

Case 2: v{a2,b1,b2,,bn2}v\in\{a_{2},b_{1},b_{2},\dots,b_{n_{2}}\}. Then d(v,c)=d(v,an11)d(v,c)=d(v,a_{n_{1}-1}) and an11,cP2a_{n_{1}-1},c\in P_{2}.

Case 3: v{a3,a4,,an122}v\in\{a_{3},a_{4},\dots,a_{\frac{n_{1}-2}{2}}\}. (This set is empty if n1=6n_{1}=6.) Then again d(v,c)=d(v,an11)d(v,c)=d(v,a_{n_{1}-1}).

Case 4: v=an12v=a_{\frac{n_{1}}{2}}. Then d(v,b1)=d(v,an11)d(v,b_{1})=d(v,a_{n_{1}-1}) and an11,b1P2a_{n_{1}-1},b_{1}\in P_{2}.

Since the other possibilities are symmetric, every metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} contains at least two vertices of Gn1,n2G_{n_{1},n_{2}}. And since every metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} contains at least n31n_{3}-1 pendant vertices attached to ii, we have dim(Gn1,n2,n3)n3+1\dim(G_{n_{1},n_{2},n_{3}})\geq n_{3}+1. Thus, by Lemma 4, dim(Gn1,n2,n3)=n3+1\dim(G_{n_{1},n_{2},n_{3}})=n_{3}+1. ∎

We now consider the counterpart of Lemma 5 for the edge metric dimension case.

Lemma 6.

Let n15n_{1}\geq 5, n21n_{2}\geq 1 and n32n_{3}\geq 2. Then edim(Gn1,n2,n3)=n3+1{\rm edim}(G_{n_{1},n_{2},n_{3}})=n_{3}+1 if n1n_{1} is odd, and edim(Gn1,n2,n3)=n3{\rm edim}(G_{n_{1},n_{2},n_{3}})=n_{3} if n1n_{1} is even.

Proof.

First assume that n1n_{1} is odd. Since n32n_{3}\geq 2, every edge metric basis contains a vertex outside of Gn1,n2G_{n_{1},n_{2}}, and by using distances, this vertex partitions E(Gn1,n2,n3)E(G_{n_{1},n_{2},n_{3}}) into sets P1={a1a2,a1an1}P_{1}=\{a_{1}a_{2},a_{1}a_{n_{1}}\}, P2={a2a3,an1an11,a2b1,an1c}P_{2}=\{a_{2}a_{3},a_{n_{1}}a_{n_{1}-1},a_{2}b_{1},a_{n_{1}}c\}, P3{a3a4,an11an12,b1b2}P_{3}\subseteq\{a_{3}a_{4},a_{n_{1}-1}a_{n_{1}-2},b_{1}b_{2}\}, etc. We show that there is no vertex in Gn1,n2,n3G_{n_{1},n_{2},n_{3}} which distinguishes edges inside these sets. Let vV(Gn1,n2,n3)v\in V(G_{n_{1},n_{2},n_{3}}). By symmetry, it suffices to distinguish the following four situations.

Case 1: v=a1v=a_{1}. Then d(v,a1a2)=d(v,a1an1)d(v,a_{1}a_{2})=d(v,a_{1}a_{n_{1}}) and a1a2,a1an1P1a_{1}a_{2},a_{1}a_{n_{1}}\in P_{1}.

Case 2: v{a2,b1,b2,,bn2}v\in\{a_{2},b_{1},b_{2},\dots,b_{n_{2}}\}. Then d(v,an1an11)=d(v,an1c)d(v,a_{n_{1}}a_{n_{1}-1})=d(v,a_{n_{1}}c) and an1an11,an1cP2a_{n_{1}}a_{n_{1}-1},a_{n_{1}}c\in P_{2}.

Case 3: v{a3,a4,,an112}v\in\{a_{3},a_{4},\dots,a_{\frac{n_{1}-1}{2}}\}. (This set is empty if n1=5n_{1}=5.) Then again d(v,an1an11)=d(v,an1c)d(v,a_{n_{1}}a_{n_{1}-1})=d(v,a_{n_{1}}c).

Case 4: v=an1+12v=a_{\frac{n_{1}+1}{2}}. Then d(v,an1an11)=d(v,a2b1)d(v,a_{n_{1}}a_{n_{1}-1})=d(v,a_{2}b_{1}) and an1an11,a2b1P2a_{n_{1}}a_{n_{1}-1},a_{2}b_{1}\in P_{2}.

Since the other possibilities are symmetric, every edge metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} contains at least two vertices of Gn1,n2G_{n_{1},n_{2}}. And since also every edge metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} contains at least n31n_{3}-1 pendant vertices attached to ii, we have edim(Gn1,n2,n3)n3+1{\rm edim}(G_{n_{1},n_{2},n_{3}})\geq n_{3}+1. By Lemma 4, we then have edim(Gn1,n2,n3)=n3+1{\rm edim}(G_{n_{1},n_{2},n_{3}})=n_{3}+1.

Now assume that n1n_{1} is even. Analogously to the proof of Lemma 4, let α=n1+12\alpha=\lfloor\frac{n_{1}+1}{2}\rfloor, γ=n112\gamma=\lfloor\frac{n_{1}-1}{2}\rfloor and δ=n12\delta=\lfloor\frac{n_{1}}{2}\rfloor. Since n1n_{1} is even, γ=δ1\gamma=\delta-1. The vertices of an edge metric basis outside Gn1,n2G_{n_{1},n_{2}} partition the edges of Gn1,n2,n3G_{n_{1},n_{2},n_{3}} into sets P1={a1a2,a1an1}P_{1}=\{a_{1}a_{2},a_{1}a_{n_{1}}\}, P2={a2a3,an1an11,a2b1,an1c}P_{2}=\{a_{2}a_{3},a_{n_{1}}a_{n_{1}-1},a_{2}b_{1},a_{n_{1}}c\}, P3{a3a4,an11an12,b1b2}P_{3}\subseteq\{a_{3}a_{4},a_{n_{1}-1}a_{n_{1}-2},b_{1}b_{2}\}, etc. For every eE(Gn1,n2,n3)e\in E(G_{n_{1},n_{2},n_{3}}), let us denote r~(e)=d(aα,e)\tilde{r}(e)=d(a_{\alpha},e). Then the values of r~\tilde{r} for the edges in P1P_{1} are γ1\gamma-1 and γ\gamma and they are different. The values of r~\tilde{r} for edges in P2P_{2} are γ2\gamma-2, δ1\delta-1, γ1\gamma-1 and δ\delta and they are different too. The values of r~\tilde{r} for edges in P3P_{3} are γ3\gamma-3, δ2\delta-2 and γ\gamma, etc. Hence, in order to identify the edges of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}, it suffices (and it is indeed necessary by the proof of Observation 3) that an edge metric generator will contain only one vertex of Gn1,n2G_{n_{1},n_{2}}, namely aαa_{\alpha}. Hence, the set {j1,j2,,jn31,aα}\{j_{1},j_{2},\dots,j_{n_{3}-1},a_{\alpha}\} is an edge metric generator of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}, see the proof of Lemma 4. Therefore, by Observation 3, we get edim(Gn1,n2,n3)=n3{\rm edim}(G_{n_{1},n_{2},n_{3}})=n_{3}. ∎

3.2 Core of the proof

To obtain the required graphs we are searching for, we will connect several copies of the graph Gn1,n2,n3G_{n_{1},n_{2},n_{3}} by adding a few edges. To this end, we need the following powerful tool.

Lemma 7.

Let G1G_{1} and G2G_{2} be two graphs which are not paths, such that for any i{1,2}i\in\{1,2\}, the graph GiG_{i} contains a metric basis SiS_{i} and an edge metric basis TiT_{i} satisfying the following conditions.

  • (1)

    There is v1S1T1v_{1}\in S_{1}\cap T_{1}.

  • (2)

    There are v2,u2S2T2v_{2},u_{2}\in S_{2}\cap T_{2} such that dG2(u2,v2)dG2(u2,z)d_{G_{2}}(u_{2},v_{2})\geq d_{G_{2}}(u_{2},z) for every zV(G2)z\in V(G_{2}).

Let GG be a graph obtained by adding the edge v1v2v_{1}v_{2} to the disjoint union of the graphs G1G_{1} and G2G_{2}. Then, dim(G)=dim(G1)+dim(G2)2\dim(G)=\dim(G_{1})+\dim(G_{2})-2 and edim(G)=edim(G1)+edim(G2)2{\rm edim}(G)={\rm edim}(G_{1})+{\rm edim}(G_{2})-2. Moreover, S=S1S2{v1,v2}S=S_{1}\cup S_{2}-\{v_{1},v_{2}\} is a metric basis of GG and T=T1T2{v1,v2}T=T_{1}\cup T_{2}-\{v_{1},v_{2}\} is an edge metric basis of GG.

Proof.

First observe that for every z2V(G2)z_{2}\in V(G_{2}), the set (S1{v1}){z2}\big{(}S_{1}-\{v_{1}\}\big{)}\cup\{z_{2}\} identifies the vertices of G1G_{1}. Analogously, if z1V(G1)z_{1}\in V(G_{1}), then the set (S2{v2}){z1}\big{(}S_{2}-\{v_{2}\}\big{)}\cup\{z_{1}\} identifies the vertices of G2G_{2}. Since G1G_{1} and G2G_{2} are not paths, |S1|2|S_{1}|\geq 2 and |S2|2|S_{2}|\geq 2. Hence, the set S=S1S2{v1,v2}S=S_{1}\cup S_{2}-\{v_{1},v_{2}\} identifies the vertices of G1G_{1} and it also identifies the vertices of G2G_{2}. Thus, to conclude that SS is a metric generator, we just may need to consider a pair of vertices x,yx,y with xV(G1)x\in V(G_{1}) and yV(G2)y\in V(G_{2}).

By (2), dG2(u2,v2)dG2(u2,z)d_{G_{2}}(u_{2},v_{2})\geq d_{G_{2}}(u_{2},z) for every zV(G2)z\in V(G_{2}). Hence, for yV(G2)y\in V(G_{2}), we have dG(u2,v2)dG(u2,y)d_{G}(u_{2},v_{2})\geq d_{G}(u_{2},y), while for xV(G1)x\in V(G_{1}) it follows dG(u2,x)dG(u2,v2)+1d_{G}(u_{2},x)\geq d_{G}(u_{2},v_{2})+1. Thus, clearly dG(u2,x)>dG(u2,y)d_{G}(u_{2},x)>d_{G}(u_{2},y). Since u2Su_{2}\in S, we conclude that SS identifies all the vertices of GG, and so it is a metric generator for GG.

Now suppose that SS^{\prime} is a metric generator for GG such that |S|<|S||S^{\prime}|<|S|. Then either |SV(G1)|<|S1|1|S^{\prime}\cap V(G_{1})|<|S_{1}|-1 or |SV(G2)|<|S2|1|S^{\prime}\cap V(G_{2})|<|S_{2}|-1. In the first case (SV(G1)){v1}(S^{\prime}\cap V(G_{1}))\cup\{v_{1}\} identifies G1G_{1} which contradicts dim(G1)=|S1|\dim(G_{1})=|S_{1}|, while in the second case (SV(G2)){v2}(S^{\prime}\cap V(G_{2}))\cup\{v_{2}\} identifies G2G_{2} which contradicts dim(G2)=|S2|\dim(G_{2})=|S_{2}|. Consequently, such a set SS^{\prime} does not exist, and we obtain that SS is a metric basis for GG, which means dim(G)=dim(G1)+dim(G2)2\dim(G)=\dim(G_{1})+\dim(G_{2})-2.

The situation for the edge metric basis is analogous. The only difference comes by noticing that we may need to consider the edge metric representation of the edge v1v2v_{1}v_{2}, that is r(v1v2|S)r(v_{1}v_{2}|S), but this is unique in GG. Observe that if z1T1{v1}z_{1}\in T_{1}-\{v_{1}\} and z2T2{v2}z_{2}\in T_{2}-\{v_{2}\}, then dG(z1,v1v2)<dG(z1,e2)d_{G}(z_{1},v_{1}v_{2})<d_{G}(z_{1},e_{2}) and dG(z2,v1v2)<dG(z2,e1)d_{G}(z_{2},v_{1}v_{2})<d_{G}(z_{2},e_{1}) for every e1E(G1)e_{1}\in E(G_{1}) and e2E(G2)e_{2}\in E(G_{2}). ∎

3.3 Conclusion of the proof

In order to complete the proof of Theorem 1, we make the following construction, that uses Lemma 7.

For some positive integer 2\ell\geq 2, we consider \ell graphs given as follows. Let G1=Gn1,n2,n3G_{1}=G_{n_{1},n_{2},n_{3}}, and let G2=G3==G=Gn1,1,2G_{2}=G_{3}=\dots=G_{\ell}=G_{n_{1},1,2}, with n15n_{1}\geq 5, n21n_{2}\geq 1 and n32n_{3}\geq 2. To distinguish vertices in distinct copies of GiG_{i}, if xx is a vertex in GkG_{k}, 1k1\leq k\leq\ell, then we denote it by xkx^{k}. Let α=n1+12\alpha=\lfloor\frac{n_{1}+1}{2}\rfloor. Then Ln1,n2,n3L^{\ell}_{n_{1},n_{2},n_{3}} is a graph obtained from the disjoint union G1G2GG_{1}\cup G_{2}\cup\dots\cup G_{\ell} by adding the edges aα1j12a^{1}_{\alpha}j^{2}_{1}, aα2j13a^{2}_{\alpha}j^{3}_{1}, …, aα1j1a^{\ell-1}_{\alpha}j^{\ell}_{1}. See Figure 3 for an example.

Refer to caption
Figure 3: The graph L7,3,43L^{3}_{7,3,4}. Note that G1G_{1} is the graph of Figure 2.

Obviously, Ln1,n2,n3L^{\ell}_{n_{1},n_{2},n_{3}} is a connected graph, and if =1\ell=1, then Ln1,n2,n3L^{\ell}_{n_{1},n_{2},n_{3}} is just Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. We next give some results concerning the metric and edge metric dimensions of such graphs.

Lemma 8.

Let n15n_{1}\geq 5, n21n_{2}\geq 1, n32n_{3}\geq 2 and 1\ell\geq 1. Then the following holds.

  • (1)

    If n1n_{1} is odd, then dim(Ln1,n2,n3)=n3\dim(L^{\ell}_{n_{1},n_{2},n_{3}})=n_{3} and edim(Ln1,n2,n3)=n3+{\rm edim}(L^{\ell}_{n_{1},n_{2},n_{3}})=n_{3}+\ell.

  • (2)

    If n1n_{1} is even, then dim(Ln1,n2,n3)=n3+\dim(L^{\ell}_{n_{1},n_{2},n_{3}})=n_{3}+\ell and edim(Ln1,n2,n3)=n3{\rm edim}(L^{\ell}_{n_{1},n_{2},n_{3}})=n_{3}.

Proof.

We only prove the result for the case when n1n_{1} is odd, since the proof for the case when n1n_{1} is even is in fact the same.

Let β=n1+32\beta=\lceil\frac{n_{1}+3}{2}\rceil. As shown in the proof of Lemma 5, dim(Gn1,n2,n3)=n3\dim(G_{n_{1},n_{2},n_{3}})=n_{3} and S={j1,j2,,jn31,aα}S=\{j_{1},j_{2},\dots,j_{n_{3}-1},a_{\alpha}\} is a metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. By Lemma 6, edim(Gn1,n2,n3)=n3+1{\rm edim}(G_{n_{1},n_{2},n_{3}})=n_{3}+1 and T={j1,j2,,jn31,aα,aβ}T=\{j_{1},j_{2},\dots,j_{n_{3}-1},a_{\alpha},a_{\beta}\} is an edge metric basis of Gn1,n2,n3G_{n_{1},n_{2},n_{3}}. We first consider the graphs G1G_{1} and G2G_{2}, in concordance with the construction of Ln1,n2,n3L^{\ell}_{n_{1},n_{2},n_{3}}. Now, for i{1,2}i\in\{1,2\}, denote the metric basis SS and an edge metric basis TT in GiG_{i} by SiS_{i} and TiT_{i}, respectively. Then aα1S1T1a_{\alpha}^{1}\in S_{1}\cap T_{1} and j12,aα2S2T2j_{1}^{2},a_{\alpha}^{2}\in S_{2}\cap T_{2}, and moreover, dG2(aα2,j1)dG2(aα2,z)d_{G_{2}}(a_{\alpha}^{2},j_{1})\geq d_{G_{2}}(a_{\alpha}^{2},z) for every zV(G2)z\in V(G_{2}). Hence, the graphs G1G_{1} and G2G_{2} satisfy the assumptions of Lemma 7. Thus, Ln1,n2,n32L^{2}_{n_{1},n_{2},n_{3}} has metric dimension n3n_{3} and edge metric dimension n3+2n_{3}+2. Moreover, S=S1S2{aα1,j12}S=S_{1}\cup S_{2}-\{a_{\alpha}^{1},j_{1}^{2}\} is a metric basis of Ln1,n2,n32L^{2}_{n_{1},n_{2},n_{3}} and T=T1T2{aα1,j12}T=T_{1}\cup T_{2}-\{a_{\alpha}^{1},j_{1}^{2}\} is an edge metric basis of Ln1,n2,n32L^{2}_{n_{1},n_{2},n_{3}}, for which aα2STa_{\alpha}^{2}\in S\cap T.

Since Ln1,n2,n32L^{2}_{n_{1},n_{2},n_{3}} and G3G_{3} satisfy the assumptions of Lemma 7, we can then proceed with Ln1,n2,n32L^{2}_{n_{1},n_{2},n_{3}} and G3G_{3} instead of G1G_{1} and G2G_{2}, and continue with this process until we reach the largest value of \ell. This concludes the proof. ∎

By using the exposition of results above, we are then able to complete the proof of Theorem 1. That is, if k1<k2k_{1}<k_{2}, then dim(L5,n2,k1k2k1)=k1\dim(L^{k_{2}-k_{1}}_{5,n_{2},k_{1}})=k_{1} and edim(L5,n2,k1k2k1)=k2{\rm edim}(L^{k_{2}-k_{1}}_{5,n_{2},k_{1}})=k_{2}, by Lemma 8. Hence, if n0=|V(L5,1,k1k2k1)|n_{0}=|V(L^{k_{2}-k_{1}}_{5,1,k_{1}})|, then for every nn0n\geq n_{0} the graph L5,1+nn0,k1k2k1L^{k_{2}-k_{1}}_{5,1+n-n_{0},k_{1}} has the required properties.

On the other hand, if k1>k2k_{1}>k_{2}, then dim(L6,n2,k1k1k2)=k1\dim(L^{k_{1}-k_{2}}_{6,n_{2},k_{1}})=k_{1} and edim(L6,n2,k2k1k2)=k2{\rm edim}(L^{k_{1}-k_{2}}_{6,n_{2},k_{2}})=k_{2}, by Lemma 8. Hence, if n0=|V(L6,1,k2k1k2)|n_{0}=|V(L^{k_{1}-k_{2}}_{6,1,k_{2}})|, then for every nn0n\geq n_{0} the graph L6,1+nn0,k2k1k2L^{k_{1}-k_{2}}_{6,1+n-n_{0},k_{2}} has the required properties. \blacksquare

4 Further work

The graphs from Figure 2, together with the graphs Ln1,n2,n3L^{\ell}_{n_{1},n_{2},n_{3}} defined above, for n1n_{1} even, allow to think that characterizing the whole class of graphs GG for which edim(G)<dim(G){\rm edim}(G)<\dim(G) is a highly challenging problem, since the structures that such graphs can have is rather wide. In this concern, observe also for instance the graphs of Figure 4, which have order 11, and other examples are the already mentioned torus graphs C4rC4tC_{4r}\square C_{4t}.

Refer to caption
Figure 4: Some graphs GG with 1111 vertices for which dim(G)>edim(G)\dim(G)>{\rm edim}(G). As in Figure 4, the squared vertices form a metric basis and the circled bolded vertices form an edge metric basis.

Notwithstanding, one could think into characterizing some special families of graphs achieving this property. Thus, some open problems that would be of interest from our point of view are the following ones.

  • Characterize the class of unicyclic graphs GG for which edim(G)<dim(G){\rm edim}(G)<\dim(G).

  • Characterize all the graphs (or maybe only the unicyclic ones) GG for which edim(G)=dim(G)1{\rm edim}(G)=\dim(G)-1.

  • Characterize all the graphs GG for which (edim(G)=2{\rm edim}(G)=2 and dim(G)=3\dim(G)=3) or (edim(G)=3{\rm edim}(G)=3 and dim(G)=4\dim(G)=4).

  • Find some necessary and/or sufficient conditions for a connected graph GG to satisfy that edim(G)<dim(G){\rm edim}(G)<\dim(G).


Acknowledgements.  The authors acknowledge partial support by Slovak research grants VEGA 1/0142/17, VEGA 1/0238/19, APVV–15–0220, APVV–17–0428, Slovenian research agency ARRS program  P1–0383 and project J1-1692.

References

  • [1] L. M. Blumenthal, Theory and applications of distance geometry, Oxford University Press, Oxford (1953).
  • [2] J. Caceres, C. Hernando, M. Mora, I. M. Pelayo, M. L. Puertas, C. Seara, and D. R. Wood, On the metric dimension of Cartesian products of graphs, SIAM J. Discrete Math. 21 (2) (2007) 423–441.
  • [3] V. Filipović, A. Kartelj, and J. Kratica, Edge metric dimension of some generalized Petersen graphs, Results Math. 74 (4) (2019) article # 182.
  • [4] J. Geneson, Metric dimension and pattern avoidance in graphs, Discrete Appl. Math. (2020). In press. DOI: 10.1016/j.dam.2020.03.001
  • [5] F. Harary and R. A. Melter, On the metric dimension of a graph, Ars Combin. 2 (1976) 191–195.
  • [6] A. Kelenc, N. Tratnik, and I. G. Yero, Uniquely identifying the edges of a graph: the edge metric dimension, Discrete Appl. Math. 251 (2018) 204–220.
  • [7] J. B. Liu, Z. Zahid, R. Nasir, and W. Nazeer, Edge version of metric dimension and doubly resolving sets of the necklace graph, Mathematics 6 (11) (2018) article # 243.
  • [8] R. Nasir, S. Zafar, and Z. Zahid, Edge metric dimension of graphs, Ars Combin. In press.
  • [9] I. Peterin and I. G. Yero, Edge metric dimension of some graph operations, Bull. Malays. Math. Sci. Soc. 43 (2020) 2465–2477.
  • [10] A. Sebő and E. Tannier, On metric generators of graphs, Math. Oper. Res. 29 (2) (2004) 383–393.
  • [11] P. J. Slater, Leaves of trees, Congr. Numer. 14 (1975) 549–559.
  • [12] Y. Zhang and S. Gao, On the edge metric dimension of convex polytopes and its related graphs, J. Comb. Optim. 39 (2) (2020) 334-350.
  • [13] E. Zhu, A. Taranenko, Z. Shao, and J. Xu, On graphs with the maximum edge metric dimension, Discrete Appl. Math. 257 (2019) 317–324.
  • [14] N. Zubrilina, On the edge dimension of a graph, Discrete Math. 341 (7) (2018) 2083–2088.