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

Characterizing the forbidden pairs for graphs to be super-edge-connected111The research is supported by National Natural Science Foundation of China (12261086).

Hazhe Ye, Yingzhi Tian222Corresponding author. E-mail: [email protected] (Y. Tian).
College of Mathematics and System Sciences, Xinjiang University, Urumqi, Xinjiang 830046, China

Abstract:

Let \mathcal{H} be a set of given connected graphs. A graph GG is said to be \mathcal{H}-free if GG contains no HH as an induced subgraph for any HH\in\mathcal{H}. The graph GG is super-edge-connected if each minimum edge-cut isolates a vertex in GG. In this paper, except for some special graphs, we characterize all forbidden subgraph sets \mathcal{H} such that every \mathcal{H}-free is super-edge-connected for ||=1|\mathcal{H}|=1 and 22.

Keywords:

Forbidden subgraphs; Super-edge-connectedness; Edge-connectivity

1 Introduction

In this paper, we consider only finite simple graphs. For notations and graph-theoretical terminology not defined here, we follow [2].

Let G=(V,E)G=(V,E) be a finite simple graph, where V=V(G)V=V(G) is the vertex set and E=E(G)E=E(G) is the edge set. For a vertex uV(G)u\in V(G), the neighborhoodneighborhood of uu in GG is NG(u)={vV(G)|vN_{G}(u)=\{v\in V(G)|\ v is adjacent to u}u\}, and the degreedegree of uu in GG is dG(u)=|NG(u)|d_{G}(u)=|N_{G}(u)|. The minimumminimum degreedegree and maximummaximum degreedegree of GG are denoted by δ(G)\delta(G) and (G)\triangle(G), respectively. For a vertex set AV(G)A\subseteq V(G), the inducedinduced subgraphsubgraph of AA in GG, denoted by G[A]G[A], is the graph with vertex set AA and two vertices uu and vv in AA are adjacent if and only if they are adjacent in GG. The distancedistance dG(u,v)d_{G}(u,v) between two vertices u,vV(G)u,v\in V(G) is the length of the shortest path between them in the graph GG.

An edge set FE(G)F\subseteq E(G) is called an edgeedge-cutcut if GFG-F is disconnected. The edgeedge-connectivityconnectivity κ(G)\kappa^{\prime}(G) of a graph GG is defined as the minimum cardinality of an edge-cut over all edge-cuts of G. The vertexvertex-connectivityconnectivity κ(G)\kappa(G) of GG can be defined similarly. The graph GG is connected if κ(G)1\kappa^{\prime}(G)\geq 1. If GG is not connected, them each maximal connected subgraph is call a componentcomponent of GG. It is well known that κ(G)κ(G)δ(G)\kappa(G)\leq\kappa^{\prime}(G)\leq\delta(G). If κ(G)=δ(G)\kappa^{\prime}(G)=\delta(G), then the graph GG is said to be maximallymaximally edgeedge-connectedconnected. In 1981, Bauer, Suffel, Boesch, and Tindell [1] proposed the concept of super-edge-connectedness. A graph GG is said to be supersuper-edgeedge-connectedconnected if each minimum edge-cut isolates a vertex of GG, that is, each minimum edge-cut of GG is the set of edges incident with some vertex in GG. Clearly, if GG is super-edge-connected, then it is also maximally edge-connected; the converse is not true, for example, the cycle CnC_{n} (n4n\geq 4) is maximally edge-connected but not super-edge-connected. There are lots of sufficient conditions for a graph to be maximally edge-connected, such as minimum degree condition [3], Ore condition [8] and diameter condition [9]. Similar sufficient conditions for a digraph to be super-edge-connected can be seen in [6]. For more results, please refer to the survey paper [7] and the references herein.

Let \mathcal{H} be a set of given connected graphs. A graph GG is said to be \mathcal{H}-freefree if there is no induced subgraph in GG isomorphic to some graph HH\in\mathcal{H}. Each graph HH in \mathcal{H} is called a forbiddenforbidden subgraphsubgraph. If ||=2|\mathcal{H}|=2, then the two graphs in \mathcal{H} are called the forbiddenforbidden pairpair. For two sets of connected graphs 1\mathcal{H}_{1} and 2\mathcal{H}_{2}, if for each graph H2H_{2} in 2\mathcal{H}_{2}, there exists a graph H1H_{1} in 1\mathcal{H}_{1} such that H1H_{1} is an induced subgraph of H2H_{2}, then denote this relation by 12\mathcal{H}_{1}\preceq\mathcal{H}_{2}. Obviously, 12\mathcal{H}_{1}\preceq\mathcal{H}_{2} implies that a 1\mathcal{H}_{1}-free graph is also 2\mathcal{H}_{2}-free. Figure 1 gives some classes of forbidden subgraphs used in this paper.

Refer to caption
Figure 1: Some classes of forbidden subgraphs.

In [5], Faudree and Gould determined the forbidden pairs for hamiltonicity of 2-connected graphs except for a finite number of exceptions. Wang, Tsuchiya and Xiong [10] characterized all the forbidden pairs R,SR,S such that every connected {R,S}\{R,S\}-free graph GG has κ(G)=κ(G)\kappa(G)=\kappa^{\prime}(G). Recently, Du, Huang and Xiong [4] characterized all the forbidden pairs R,SR,S such that every connected {R,S}\{R,S\}-free graph GG is maximally edge-connected.

Theorem 1.1.

([10]) Let SS be a connected graph. Then GG being a connected SS-free graph implies κ(G)=κ(G)\kappa(G)=\kappa^{\prime}(G) if and only if SS is an induced subgraph of P3P_{3}.

Theorem 1.2.

([10]) Let \mathcal{H} be a set of two connected graphs such that each member of \mathcal{H} is not an induced subgraph of P3P_{3}. Then GG being a connected \mathcal{H}-free graph implies κ(G)=κ(G)\kappa(G)=\kappa^{\prime}(G) if and only if {Z1,P5},{Z1,K1,4},{Z1,T1,1,2}\mathcal{H}\preceq\left\{Z_{1},P_{5}\right\},\mathcal{H}\preceq\left\{Z_{1},K_{1,4}\right\},\mathcal{H}\preceq\left\{Z_{1},T_{1,1,2}\right\}, {H0,P4}\mathcal{H}\preceq\left\{H_{0},P_{4}\right\}, or {H0,K1,3}\mathcal{H}\preceq\left\{H_{0},K_{1,3}\right\}.

Theorem 1.3.

([4]) Let SS be a connected graph. Then GG being a connected S-free graph implies κ(G)=\kappa^{\prime}(G)= δ(G)\delta(G) if and only if SS is an induced subgraph of P4P_{4}.

Theorem 1.4.

([4]) Let ={R,S}\mathcal{H}=\{R,S\} be a set of two connected graphs such that both RR and SS are not an induced subgraph of P4P_{4}. Then GG being a connected \mathcal{H}-free graph implies κ(G)=δ(G)\kappa^{\prime}(G)=\delta(G) if and only if {H1,P5},{Z2,P6}\mathcal{H}\preceq\left\{H_{1},P_{5}\right\},\mathcal{H}\preceq\left\{Z_{2},P_{6}\right\}, or {Z2,T1,1,3}\mathcal{H}\preceq\left\{Z_{2},T_{1,1,3}\right\}.

Motivated by the results above, we will characterize the forbidden pairs for a graph to be super-edge-connected in this paper. In the next section, the main results will be presented. The proof of the main theorem will be given in the last section.

2 Main results

Theorem 2.1.

Let SS be a connected graph. Then GG being a connected S-free graph implies GG is super-edge-connected if and only if SS is an induced subgraph of P3P_{3}.

Proof.

Let GG be a connected P3P_{3}-free graph. Then GG is a complete graph, and thus GG is super-edge-connected.

Let SS be a connected graph such that every connected SS-free graph is super-edge-connected. Since GiG_{i} in Figure 2 is not super-edge-connected for i{1,,6}i\in\{1,\cdots,6\}, then we know GiG_{i} is not SS-free and contains SS as an induced subgraph. We observe that the common induced subgraph of the graphs in G3{G}_{3} and G6{G}_{6} is a path and the longest induced path of the graphs in G1{G}_{1} is P3P_{3}, then SS must be an induced subgraph of P3P_{3}. The proof is complete. ∎

The CartesianCartesian productproduct of two graphs G1G_{1} and G2G_{2}, denoted by G1G2G_{1}\square G_{2}, is defined on the vertex sets V(G1)×V(G2)V(G_{1})\times V(G_{2}), and (x1,y1)(x2,y2)(x_{1},y_{1})(x_{2},y_{2}) is an edge in G1G2G_{1}\square G_{2} if and only if one of the following is true: (ii) x1=x2x_{1}=x_{2} and y1y2E(G2)y_{1}y_{2}\in E(G_{2}); (iiii) y1=y2y_{1}=y_{2} and x1x2E(G1)x_{1}x_{2}\in E(G_{1}).

In the following, we try to characterize the forbidden pairs for a graph to be super-edge-connected. By Theorem 2.1, a connected P3P_{3}-free graph is super-edge-connected. Thus for any connected graph SS, we obtain that if GG is {P3,S}\left\{P_{3},S\right\}-free, then GG is super-edge-connected. So we assume that both RR and SS are not an induced subgraph of P3P_{3} in the following main theorem in this paper.

Theorem 2.2.

Let ={R,S}\mathcal{H}=\{R,S\} be a set of two connected graphs such that both RR and SS are not an induced subgraph of P3P_{3}. Then GG being a connected \mathcal{H}-free graph implies GG is super-edge-connected if and only if (i) GC4G\ncong C_{4} and {H0,P4}\mathcal{H}\preceq\left\{H_{0},P_{4}\right\}, or (ii) GP2P3G\ncong P_{2}\Box P_{3}, GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4) and {Z1,T1,1,2}\mathcal{H}\preceq\left\{Z_{1},T_{1,1,2}\right\}.

3 The Proof of Theorem 2.2

In Figure 2, we construct some classes of non super-edge-connected graphs GiG_{i} (where i{1,,6}i\in\{1,\cdots,6\}) on nn vertices, which will be used in the proof of the main result. To distinguish these graphs, we assume n8n\geq 8. CnC_{n} denotes the cycle on nn vertices and KnK_{n} denotes the complete graph on nn vertices.

Refer to caption
Figure 2: Some classes of non super-edge-connected graphs.
The Proof of Theorem 2.2..

We first prove the necessity. Assume ={R,S}\mathcal{H}=\{R,S\} is a set of connected graphs such that both RR and SS are not an induced subgraph of P3P_{3}, and every connected \mathcal{H}-free graph is super-edge-connected. Let GG be a connected \mathcal{H}-free graph. Then GP2P3G\ncong P_{2}\Box P_{3} and GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4). Since GiG_{i} in Figure 2 is not super-edge-connected, we obtain that GiG_{i} is not \mathcal{H}-free and contains at least one of RR and SS as an induced subgraph for i{1,,6}i\in\{1,\cdots,6\}.

Claim 1. Either RR or SS is an induced subgraph of H0H_{0}.

By contradiction, assume that neither RR nor SS is an induced subgraph of H0H_{0}. Without loss of generality, assume that RR is an induced subgraph of G1G_{1}. Since RR is not an induced subgraph of H0H_{0}. We observe that RR must contain a KtK_{t} as an induced subgraph with t4t\geq 4 . For i{2,3,4,5,6}i\in\{2,3,4,5,6\}, GiG_{i} is RR-free. Then they contain SS as an induced subgraph. We observe that the common induced subgraph of the graphs in G3{G}_{3} and G6{G}_{6} is a path and the longest induced path of G2G_{2} is P3P_{3}, then SS should be an induced subgraph of P3P_{3}, a contradiction. Claim 1 is thus proved.

By Claim 1, assume, without loss of generality, that RR is an induced subgraph of H0H_{0}. We distinguish two cases as follows.

Case 1. RR is H0H_{0}.

For i{3,4,5,6}i\in\{3,4,5,6\}, GiG_{i} is RR-free. Then they contain SS as an induced subgraph. We observe that the common induced subgraph of the graphs in G3{G}_{3} and G6{G}_{6} is a path and the longest induced path of G4G_{4} is P4P_{4}, then SS should be an induced subgraph of P4P_{4}. So ={R,S}{H0,P4}\mathcal{H}=\{R,S\}\preceq\left\{H_{0},P_{4}\right\} .

Case 2. RR is an induced subgraph of Z1Z_{1}.

Since RR is not an induced subgraph of P3P_{3}, we get that RR must contain a triangle. For i{3,4,5}i\in\{3,4,5\}, GiG_{i} is RR-free. Then they contain SS as an induced subgraph. We observe that the maximal common induced subgraph of the graphs in G4{G}_{4} and G5{G}_{5} is a T1,1,2T_{1,1,2}, then SS should be an induced subgraph of T1,1,2T_{1,1,2}. So ={R,S}{Z1,T1,1,2}\mathcal{H}=\{R,S\}\preceq\left\{Z_{1},T_{1,1,2}\right\}.

Now we are going to prove the sufficiency. By contradiction, assume GG is a connected \mathcal{H}-free graph, but GG is not super-edge-connected, where (ii) GC4G\ncong C_{4} and {H0,P4}\mathcal{H}\preceq\left\{H_{0},P_{4}\right\}, or (iiii) GP2P3G\ncong P_{2}\Box P_{3}, GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4) and {Z1,T1,1,2}\mathcal{H}\preceq\left\{Z_{1},T_{1,1,2}\right\}. Since GG is not super-edge-connected, there is a minimum edge-cut FF such that GFG-F has no isolated vertices. Clearly, GFG-F has only two components, say XX and YY. Let X1=V(X)V(F)X_{1}=V\left(X\right)\cap V(F) and Y1=V(Y)V(F)Y_{1}=V\left(Y\right)\cap V(F). Denote X2=V(X)X1X_{2}=V(X)\setminus X_{1} and Y2=V(Y)Y1Y_{2}=V(Y)\setminus Y_{1}. Note that |V(X)|,|V(Y)|2|V(X)|,|V(Y)|\geq 2 and |X1|,|Y1||F|=κ(G)δ(G)|X_{1}|,|Y_{1}|\leq|F|=\kappa^{\prime}(G)\leq\delta(G).

Claim 2. If X2=X_{2}=\emptyset, then XX is a complete graph with order δ(G)\delta(G), each vertex in XX is incident with exactly one edge in FF, and |X|=|X1|=|F|=κ(G)=δ(G)|X|=|X_{1}|=|F|=\kappa^{\prime}(G)=\delta(G); similarly, if Y2=Y_{2}=\emptyset, then YY is a complete graph with order δ(G)\delta(G), each vertex in YY is incident with exactly one edge in FF, and |Y|=|Y1|=|F|=κ(G)=δ(G)|Y|=|Y_{1}|=|F|=\kappa^{\prime}(G)=\delta(G).

By κ(G)δ(G)\kappa^{\prime}(G)\leq\delta(G), |X1||V(X)||X_{1}|\leq|V(X)| and |X1||F|=κ(G)|X_{1}|\leq|F|=\kappa^{\prime}(G), we have the following inequalities.

|E(X)|\displaystyle\left|E\left(X\right)\right| =12(xV(X)dG(x)κ(G))\displaystyle=\frac{1}{2}\left(\sum_{x\in V\left(X\right)}d_{G}(x)-\kappa^{\prime}(G)\right)
12(δ(G)|V(X)|κ(G))\displaystyle\geq\frac{1}{2}\left(\delta(G)\left|V\left(X\right)\right|-\kappa^{\prime}(G)\right)
12(δ(G)|X1|κ(G))\displaystyle\geq\frac{1}{2}\left(\delta(G)|X_{1}|-\kappa^{\prime}(G)\right)
12κ(G)(|X1|1)\displaystyle\geq\frac{1}{2}\kappa^{\prime}(G)\left(|X_{1}|-1\right)
12|X1|(|X1|1).\displaystyle\geq\frac{1}{2}|X_{1}|\left(|X_{1}|-1\right).

If X2=X_{2}=\emptyset, then all the inequalities above will be equalities. Thus we obtain that XX is a complete graph with order δ(G)\delta(G), each vertex in XX is incident with exactly one edge in FF, and |X|=|X1|=|F|=κ(G)=δ(G)|X|=|X_{1}|=|F|=\kappa^{\prime}(G)=\delta(G).

By Claim 2, we consider two cases in the following.

Case 1. At least one of X2X_{2} and Y2Y_{2} is an empty set.

Assume, without loss of generality, that X2=X_{2}=\emptyset. Then, by Claim 2, XX is a complete graph with order δ(G)\delta(G), each vertex in XX is incident with exactly one edge in FF, and |X|=|X1|=|F|=κ(G)=δ(G)|X|=|X_{1}|=|F|=\kappa^{\prime}(G)=\delta(G). In addition, by |V(X)|2|V(X)|\geq 2, we have δ(G)2\delta(G)\geq 2.

Subcase 1.1. GC4G\ncong C_{4} and {H0,P4}\mathcal{H}\preceq\left\{H_{0},P_{4}\right\}.

Suppose Y2=Y_{2}=\emptyset. Then GG is the union of two complete graphs on δ(G)\delta(G) vertices, together with the perfect matching FF, that is, GG is isomorphic to the Cartesian product graph Kδ(G)K2K_{\delta(G)}\Box K_{2}. If δ(G)3\delta(G)\geq 3, then there exists an induced P4P_{4}, a contradiction. Otherwise, if δ(G)=2\delta(G)=2, then GC4G\cong C_{4}, also a contradiction.

Suppose Y2Y_{2}\neq\emptyset. Then by YY is connected, there is a path x1x2y1y2x_{1}x_{2}y_{1}y_{2}, where x1,x_{1}, x2V(X)x_{2}\in V\left(X\right), y1Y1y_{1}\in Y_{1} and y2Y2y_{2}\in Y_{2}. If x1y1E(G)x_{1}y_{1}\notin E(G), then x1x2y1y2x_{1}x_{2}y_{1}y_{2} is an induced path of GG, a contradiction. Thus we assume x1y1E(G)x_{1}y_{1}\in E(G). Since dG(y2)δ(G)2d_{G}\left(y_{2}\right)\geq\delta(G)\geq 2, y2y_{2} has a neighbor y3y_{3} different from y1y_{1}. If y1y3E(G)y_{1}y_{3}\notin E(G), then G[{x1,y1,y2,y3}]P4G\left[\{x_{1},y_{1},y_{2},y_{3}\}\right]\cong P_{4}, a contradiction. Otherwise, if y1y3E(G)y_{1}y_{3}\in E(G), then G[{x1,x2,y1,y2,y3}]H0G\left[\{x_{1},x_{2},y_{1},y_{2},y_{3}\}\right]\cong H_{0}, also a contradiction.

Subcase 1.2. GP2P3G\ncong P_{2}\Box P_{3}, GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4) and {Z1,T1,1,2}\mathcal{H}\preceq\left\{Z_{1},T_{1,1,2}\right\}.

Clearly, |X1|=|F||Y1||X_{1}|=|F|\geq|Y_{1}|. If |X1|>|Y1||X_{1}|>|Y_{1}|, then there exists a vertex in Y1Y_{1}, say y1y_{1}, has at least two neighbors in X1X_{1}, say x1x_{1} and x2x_{2}. By YY is connected, y1y_{1} has at least one neighbor, say y2y_{2}, in V(Y)V(Y). Since XX is a complete graph and each vertex in XX is incident with exactly one edge in FF, we know that G[{x1,x2,y1,y2}]Z1G\left[\{x_{1},x_{2},y_{1},y_{2}\}\right]\cong Z_{1}, a contradiction. Thus, we assume |X1|=|Y1||X_{1}|=|Y_{1}| in the following.

Since |X1|=|Y1|=|F||X_{1}|=|Y_{1}|=|F|, we know that each vertex in X1X_{1} has only one neighbor in Y1Y_{1} and each vertex in Y1Y_{1} has only one neighbor in X1X_{1}. If |X1|=|Y1|3|X_{1}|=|Y_{1}|\geq 3, then the induced subgraph by any three vertex in X1X_{1}, say x1,x2,x3x_{1},x_{2},x_{3}, together with the neighbor of x1x_{1} in Y1Y_{1} is isomorphic to Z1Z_{1}, a contradiction. It is remaining to consider |X1|=|Y1|=2|X_{1}|=|Y_{1}|=2. Assume X1={x1,x2}X_{1}=\{x_{1},x_{2}\}, Y1={y1,y2}Y_{1}=\{y_{1},y_{2}\} and F={x1y1,x2y2}F=\{x_{1}y_{1},x_{2}y_{2}\}.

Suppose y1y2E(G)y_{1}y_{2}\in E(G). Since GG is not a cycle, y1y_{1} or y2y_{2}, say y2y_{2} has a neighbor y3y_{3} in Y2Y_{2}. If y1y3E(G)y_{1}y_{3}\in E(G), then G[{x1,y1,y2,y3}]Z1G[\{x_{1},y_{1},y_{2},y_{3}\}]\cong Z_{1}, a contradiction. So y1y3E(G)y_{1}y_{3}\notin E(G). Since δ(G)2\delta(G)\geq 2, y3y_{3} has another neighbor y4y_{4} different from y2y_{2}. If y2y4E(G)y_{2}y_{4}\in E(G), then G[{x2,y2,y3,y4}]Z1G[\{x_{2},y_{2},y_{3},y_{4}\}]\cong Z_{1}, a contradiction. So y2y4E(G)y_{2}y_{4}\notin E(G). If y1y4E(G)y_{1}y_{4}\notin E(G), then G[{x2,y1,y2,y3,y4}]T1,1,2G[\{x_{2},y_{1},y_{2},y_{3},y_{4}\}]\cong T_{1,1,2}, a contradiction. So y1y4E(G)y_{1}y_{4}\in E(G). If V(G)={x1,x2,y1,y2,y3,y4}V(G)=\{x_{1},x_{2},y_{1},y_{2},y_{3},y_{4}\}, then GP2P3G\cong P_{2}\Box P_{3}, a contradiction. Thus, by GG is connected, there is a vertex y5y_{5} adjacent to y1y_{1}, y2y_{2}, y3y_{3}, or y4y_{4}. We only consider the case y1y5E(G)y_{1}y_{5}\in E(G) here, others can be analysed similarly. If y4y5E(G)y_{4}y_{5}\in E(G), then G[{x1,y1,y4,y5}]Z1G[\{x_{1},y_{1},y_{4},y_{5}\}]\cong Z_{1}, a contradiction. Otherwise, if y4y5E(G)y_{4}y_{5}\notin E(G), then G[{x1,x2,y1,y4,y5}]T1,1,2G[\{x_{1},x_{2},y_{1},y_{4},y_{5}\}]\cong T_{1,1,2}, also a contradiction.

Suppose y1y2E(G)y_{1}y_{2}\notin E(G). Since YY is connected, there is a path z1z2ztz_{1}z_{2}\cdots z_{t} connecting y1y_{1} and y2y_{2} in YY, where z1=y1z_{1}=y_{1}, zt=y2z_{t}=y_{2} and t3t\geq 3. Furthermore, we choose this path z1z2ztz_{1}z_{2}\cdots z_{t} to be a shortest path connecting y1y_{1} and y2y_{2} in YY. Since GG is not a cycle, there is a vertex wY{z1,z2,,zt}w\in Y\setminus\{z_{1},z_{2},\cdots,z_{t}\} such that ww is adjacent to some vertex in {z1,z2,,zt}\{z_{1},z_{2},\cdots,z_{t}\}. Assume wziE(G)wz_{i}\in E(G) and wzjE(G)wz_{j}\notin E(G) for j<ij<i. Let z0=x1z_{0}=x_{1} and z1=x2z_{-1}=x_{2} . If wzi+1E(G)wz_{i+1}\in E(G), then G[{w,zi1,zi,zi+1}]Z1G[\{w,z_{i-1},z_{i},z_{i+1}\}]\cong Z_{1}, a contradiction. Otherwise, if wzi+1E(G)wz_{i+1}\notin E(G), then G[{w,zi2,zi1,zi,zi+1}]T1,1,2G[\{w,z_{i-2},z_{i-1},z_{i},z_{i+1}\}]\cong T_{1,1,2}, also a contradiction.

Case 2. Neither X2X_{2} nor Y2Y_{2} is an empty set.

Since the distance between a vertex in X2X_{2} and a vertex in Y2Y_{2} is at least three, GG must contains P4P_{4} as an induced path. So the case GC4G\ncong C_{4} and {H0,P4}\mathcal{H}\preceq\left\{H_{0},P_{4}\right\} can not happen. Thus we only consider the case that GP2P3G\ncong P_{2}\Box P_{3}, GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4) and {Z1,T1,1,2}\mathcal{H}\preceq\left\{Z_{1},T_{1,1,2}\right\} in the following. We distinguish two subcases as follows.

Subcase 2.1. GG contains a path x2x1y1y2x_{2}x_{1}y_{1}y_{2}, where xiXix_{i}\in X_{i} and yiYiy_{i}\in Y_{i} for i=1,2i=1,2.

Subcase 2.1.1. |NG(x2)X2|2\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|\geq 2 or |NG(y2)Y2|2\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|\geq 2.

Assume |NG(x2)X2|2\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|\geq 2. Then there exist two vertices x2x_{2}^{\prime}, x2′′X2x_{2}^{\prime\prime}\in X_{2} such that x2x2x_{2}x_{2}^{\prime}, x2x2′′E(G)x_{2}x_{2}^{\prime\prime}\in E(G). If x2x2′′x_{2}^{\prime}x_{2}^{\prime\prime}, x2x1x_{2}^{\prime}x_{1}, x2′′x1E(G)x_{2}^{\prime\prime}x_{1}\notin E(G), then G[{x1,x2,x2,x2′′,y1}]T1,1,2G\left[\{x_{1},x_{2},x_{2}^{\prime},x_{2}^{\prime\prime},y_{1}\}\right]\cong T_{1,1,2}, a contradiction. If at least one of these three edges exists, then GG contains an induced subgraph isomorphic to Z1Z_{1}, also a contradiction. For example, if x2x1E(G)x_{2}^{\prime}x_{1}\in E(G), then G[{x1,x2,x2,y1}]Z1G\left[\{x_{1},x_{2},x_{2}^{\prime},y_{1}\}\right]\cong Z_{1}. The case |NG(y2)Y2|2\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|\geq 2 can be proved similarly.

Subcase 2.1.2. |NG(x2)X2|=1\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|=1 and |NG(y2)Y2|=1\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|=1.

Let x3NG(x2)X2x_{3}\in N_{G}\left(x_{2}\right)\cap X_{2} and y3NG(y2)Y2y_{3}\in N_{G}\left(y_{2}\right)\cap Y_{2}. If x1x3x_{1}x_{3} or y1y3E(G)y_{1}y_{3}\in E(G), then GG contains an induced Z1Z_{1}. For example, if x1x3E(G)x_{1}x_{3}\in E(G), then G[{x1,x2,x3,y1}]Z1G\left[\{x_{1},x_{2},x_{3},y_{1}\}\right]\cong Z_{1}. Hence, x1x3E(G)x_{1}x_{3}\notin E(G) and y1y3E(G)y_{1}y_{3}\notin E(G).

Suppose dG(x1)3d_{G}\left(x_{1}\right)\geq 3. Let wNG(x1){x2,y1}w\in N_{G}\left(x_{1}\right)\setminus\{x_{2},y_{1}\}. If wX2w\in X_{2}, then G[{x1,x2,y1,w}]Z1G\left[\{x_{1},x_{2},y_{1},w\}\right]\cong Z_{1} when wx2E(G)wx_{2}\in E(G), and G[{x1,x2,y1,y2,w}]T1,1,2G\left[\{x_{1},x_{2},y_{1},y_{2},w\}\right]\cong T_{1,1,2} when wx2E(G)wx_{2}\notin E(G), contradicts to the assumption. By a similar argument, we can obtain contradictions for wX1w\in X_{1} and wY1w\in Y_{1}. Thus dG(x1)=2d_{G}\left(x_{1}\right)=2. Similarly, dG(y1)=2d_{G}\left(y_{1}\right)=2. Since GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4), then there exists a vertex in GG such that its degree is at least three. Assume, without loss of generality, V(X)V(X) has a vertex with degree at least three. Choose such a vertex z1z_{1} such that its distance to x1x_{1} is minimum in the component XX. Let z1z2zr(zr=x1)z_{1}z_{2}\cdots z_{r}(z_{r}=x_{1}) be the shortest path between z1z_{1} and x1x_{1}. Then dG(zi)=2d_{G}(z_{i})=2 for 2ir2\leq i\leq r. Let z1z_{1}^{\prime}, z1′′NG(z1){z2}z_{1}^{\prime\prime}\in N_{G}\left(z_{1}\right)\setminus\{z_{2}\}. Then we can find an induced Z1Z_{1} or T1,1,2T_{1,1,2} according to z1z1′′z_{1}^{\prime}z_{1}^{\prime\prime} is an edge in GG or not.

Subcase 2.1.3. |NG(x2)X2|=0\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|=0 and |NG(y2)Y2|1\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|\leq 1, or |NG(x2)X2|1\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|\leq 1 and |NG(y2)Y2|=0\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|=0.

Without loss of generality, assume that|NG(x2)X2|=0\left|N_{G}\left(x_{2}\right)\cap X_{2}\right|=0 and |NG(y2)Y2|1\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|\leq 1. Since NG(x2)X1N_{G}\left(x_{2}\right)\subseteq X_{1} and |X1||F|=κ(G)δ(G)|X_{1}|\leq|F|=\kappa^{\prime}(G)\leq\delta(G), we have dG(x2)=|X1|=|F|=κ(G)=δ(G)d_{G}\left(x_{2}\right)=|X_{1}|=|F|=\kappa^{\prime}(G)=\delta(G). Note that |Y1||F|=|X1||Y_{1}|\leq|F|=|X_{1}|.

Suppose |X1|>|Y1||X_{1}|>|Y_{1}|. Then |NG(y2)Y2|1\left|N_{G}\left(y_{2}\right)\cap Y_{2}\right|\geq 1. Let y3NG(y2)Y2y_{3}\in N_{G}\left(y_{2}\right)\cap Y_{2}. If y1y3E(G)y_{1}y_{3}\in E(G), then G[{x1,y1,y2,y3}]Z1G\left[\{x_{1},y_{1},y_{2},y_{3}\}\right]\cong Z_{1}, a contradiction. So assume y1y3E(G)y_{1}y_{3}\notin E(G). If |NG(y1)X1|2\left|N_{G}\left(y_{1}\right)\cap X_{1}\right|\geq 2, then we can find an induced Z1Z_{1} or T1,1,2T_{1,1,2} according to any two vertices in NG(y1)X1N_{G}\left(y_{1}\right)\cap X_{1} are adjacent or not. So assume y1y_{1} has exactly one neighbor x1x_{1} in X1X_{1}. By |X1|>|Y1|1|X_{1}|>|Y_{1}|\geq 1, we get |X1|2|X_{1}|\geq 2. Since y1y_{1} has exactly one neighbor x1x_{1} in X1X_{1}, we have |Y1|2|Y_{1}|\geq 2, thus |X1|3|X_{1}|\geq 3. If there is at least one edge in G[X1]G[X_{1}], then we can find an induced Z1Z_{1}, a contradiction. Otherwise, we can find an induced T1,1,2T_{1,1,2} rooted at x2x_{2}, also a contradiction.

Suppose |X1|=|Y1||X_{1}|=|Y_{1}|. Then by |X1|=|F|=κ(G)=δ(G)|X_{1}|=|F|=\kappa^{\prime}(G)=\delta(G), we know that each vertex in X1X_{1} has only one neighbor in Y1Y_{1} and each vertex in Y1Y_{1} has only one neighbor in X1X_{1}. That is, FF is perfect matching in G[X1Y1]G[X_{1}\cup Y_{1}] connecting X1X_{1} and Y1Y_{1}.

Suppose |X1|=|Y1|3|X_{1}|=|Y_{1}|\geq 3. Then according to there is an edge in G[X1]G[X_{1}] or not, we can find an induced Z1Z_{1} or T1,1,2T_{1,1,2}, a contradiction. Suppose |X1|=|Y1|=2|X_{1}|=|Y_{1}|=2. By a similar argument as Subcase 1.2, we can obtain a contradiction. Suppose |X1|=|Y1|=1|X_{1}|=|Y_{1}|=1. Since GPn,Cn(n4)G\ncong P_{n},C_{n}\ (n\geq 4), then there exists a vertex in GG such that its degree is at least three. By a similar argument as Subcase 2.1.2, we can also obtain a contradiction.

Subcase 2.2. GG contains no path x2x1y1y2x_{2}x_{1}y_{1}y_{2} satisfying xiXix_{i}\in X_{i} and yiYiy_{i}\in Y_{i} for i=1,2i=1,2.

Let X11={xX1:NG(x)X2}X_{1}^{1}=\left\{x\in X_{1}:N_{G}(x)\cap X_{2}\neq\emptyset\right\} and Y11={yY1:NG(y)Y2}Y_{1}^{1}=\left\{y\in Y_{1}:N_{G}(y)\cap Y_{2}\neq\emptyset\right\}. Denote X12=X1X11X_{1}^{2}=X_{1}-X_{1}^{1} and Y12=Y1Y11Y_{1}^{2}=Y_{1}-Y_{1}^{1}. Since GG contains no path x2x1y1y2x_{2}x_{1}y_{1}y_{2} satisfying xiXix_{i}\in X_{i} and yiYiy_{i}\in Y_{i} for i=1,2i=1,2, we obtain that X12X_{1}^{2}\neq\emptyset, Y12Y_{1}^{2}\neq\emptyset and there is no edge connecting X11X_{1}^{1} and Y11Y_{1}^{1}. Since XX is connected, there is an edge connecting X11X_{1}^{1} and X12X_{1}^{2}, say x11x12E(G)x_{1}^{1}x_{1}^{2}\in E(G), where x11X11x_{1}^{1}\in X_{1}^{1} and x12X12x_{1}^{2}\in X_{1}^{2}. Let x2x_{2} be a neighbor of x11x_{1}^{1} in X2X_{2} and y1y_{1} be a neighbor of x11x_{1}^{1} in Y1Y_{1}. Since |X11|<δ(G)|X_{1}^{1}|<\delta(G), x2x_{2} has a neighbor, say x3x_{3}, in X2X_{2}. If x11x3E(G)x_{1}^{1}x_{3}\in E(G), then G[{x11,x2,x3,y1}]Z1G\left[\{x_{1}^{1},x_{2},x_{3},y_{1}\}\right]\cong Z_{1}, a contradiction. So assume x11x3E(G)x_{1}^{1}x_{3}\notin E(G). If x12y1E(G)x_{1}^{2}y_{1}\in E(G), then G[{x11,x12,x2,y1}]Z1G\left[\{x_{1}^{1},x_{1}^{2},x_{2},y_{1}\}\right]\cong Z_{1}, a contradiction. So assume x12y1E(G)x_{1}^{2}y_{1}\notin E(G). Thus G[{x11,x12,x2,x3,y1}]T1,1,2G\left[\{x_{1}^{1},x_{1}^{2},x_{2},x_{3},y_{1}\}\right]\cong T_{1,1,2}, also a contradiction.

Since all cases lead to contradictions, the proof is thus complete. ∎

References

  • [1] D. Bauer, C. Suffel, F. Boesch, R. Tindell, Connectivity extremal problems and the design of reliable probabilistic networks, in: The Theory and Application of Graphs, Kalamazoo MI, 1980, Wiley, New York, 1981, pp. 45-54.
  • [2] J. A. Bondy, U. S. R. Murty, Graph Theory, Graduate Texts in Mathematics 244, Springer, Berlin, 2008.
  • [3] G. Chartrand, A graph-theoretic approach to a communications problem, SIAM J. Appl. Math. 14 (1966) 778-781.
  • [4] J. Du, Z. Huang, L. Xiong, Characterizing forbidden pairs for the edge-connectivity of a connected graph to be its minimum degree, Axioms 11 (2022) 219.
  • [5] R. J. Faudree, R. J. Gould, Characterizing forbidden pairs for Hamiltonian properties, Discrete Math. 173 (1997) 45-60.
  • [6] M. A. Fiol, On super-edge-connected digraphs and bipartite digraphs, J. Graph Theory 16 (1992) 545-555.
  • [7] A. Hellwig, L. Volkmann, Maximally edge-connected and vertex-connected graphs and digraphs: A survey, Discrete Math. 308 (2008) 3265-3296.
  • [8] L. Lesniak, Results on the edge-connectivity of graphs, Discrete Math. 8 (1974) 351-354.
  • [9] J. Plesnik, Critical graphs of given diameter, Acta Fac. Rerum Natur. Univ. Commenian Math. 30 (1975) 71-93.
  • [10] S. Wang, S. Tsuchiya, L. Xiong, Forbidden pairs for equality of connectivity and edge-connectivity of graphs, Graphs Comb. 35 (2019) 419-426.