1. Introduction
In this paper, we consider finite simple graphs only. For terminology and notation not defined here, we refer the readers to Bondy and Murty [
1].
Let G be a connected graph with vertex set and edge set . Therefore, and mean the order, size, connectivity, edge-connectivity and minimum degree of G, respectively. Suppose u is a vertex of G. Then denotes , which is also called the neighbors of u in the graph G. Let (or respectively). The subgraph of G induced by S is denoted by , vertex induced subgraph and edge induced subgrph. Furthermore, we use to denote the subgraph (or respectively). The distance between two vertices , denoted by , is the length of a shortest path between the two vertices x and y, while the diameter of a graph G, denoted by , is the greatest distance between any pair of vertices of G.
Let H be a given graph. A graph G is said to be H-free if any induced subgraph of G is not isomorphic to H. If G is H-free, then H is called a forbidden subgraph of G. Note that if is an induced subgraph of , then every -free graph is also -free. Let be a set of connected graphs, the graph G is -free if G is H-free for every . For two sets and of connected graphs, we write which means that for every graph , there exists a graph such that is an induced subgraph of . By the definition, we know that if , then clearly every -free graph is also -free.
It always means that we use
,
to denote the complete graph of order
n, and the complete bipartite graph with partition sets of size
s and
t, respectively. So the
is a trivial graph,
is a triangle,
is a star (the
is also called a
claw). A clique
C is a subgraph of a graph
G such that
is a complete graph, and the
clique number of a graph
G is the maximum cardinality of a clique of
G. Then we will show some special graphs which are needed: (see
Figure 1)
, the path with order i, hence and ;
, a graph obtained by identifying a vertex of a with an end-vertex of a ;
, a graph obtained by identifying a vertex of a with the one-degree vertex of a ;
, a graph consisting of three paths and with the common starting vertex.
Let
X and
Y be nonempty subsets of
. We denote by
the set of edges of
G with one end in
X and the other end in
Y, and by
their number. When
, the set
is called the
edge cut of
G associated with
X. An edge cut set
S with the minimum number of edges is called a
minimum edge cut. It is well-known that
. In [
2], Wang, Tsuchiya and Xiong characterize all the pairs
such that every connected
-free graph
G has
.
Theorem 1 (Wang et al. [
2]).
Let S be a connected graph. Then G is a connected S-free graph which implies if and only if S is an induced subgraph of . Theorem 2 (Wang et al. [
2]).
Let be a set of two connected graphs such that . Then G is a connected -free graph that if and only if , , , or , where is the unique simple graph with degree sequence 4, 2, 2, 2, 2, i.e., which can be obtained from by the contracted edge whose endvertices are of degree 3. In [
3], Hellwig and Volkmann introduce several sufficient conditions for
.
Theorem 3. Let G be a connected graph satisfying one of the following conditions:
- 1.
(Chartrand [
4])
,- 2.
(Lesniak [
5])
for all pairs of nonadjacent vertices,- 3.
(Plesník [
6])
,- 4.
(Volkmann [
7])
G is bipartite and ,- 5.
(Plesník and Znám [
8])
there are no four vertices with,
- 6.
(Plesník and Znám [
8])
G is bipartite and ,- 7.
(Xu [
9])
there exist pairs of vertices such that for ,- 8.
(Dankelmann and Volkmann [
10])
and .
Then .
2. Our Results
In this paper, we characterize all forbidden subgraphs sets of graphs such that every connected -free graph implies for .
Theorem 4. Let S be a connected graph. Then G being a connected S-free graph implies if and only if S is an induced subgraph of .
Theorem 5. Let be a set of two connected graphs such that R and S are not an induced subgraph of . Then G being a connected -free graph implies if and only if , , or .
Note that whenever a connected graph G satisfies or , it satisfies . Then we want to characterize the forbidden subgraphs for .
In fact, for are three invariants of G with , we also present a general result which may help us to deal with the relationship between them. In order to state the result clearly, we further introduce some notations. For two sets of given graphs and , we use to denote the set with order n, in which each element is the common induced subgraph of one graph in and one graph in , respectively, i.e., is the common induced subgraph of and .
Now we may get the following result (A similar result by replacing the parameter with a set of subgraphs; you may see [
11], in its last section).
Theorem 6. Let G be a connected graph, and are three invariants of G with . If the following statements hold:
- 1.
G is -free implies if and only if ;
- 2.
G is -free implies if and only if ,
then G is -free implies if and only if .
Proof. First suppose G is -free and , then for some and . By the definition of , we can see that and . Therefore, G is -free and -free. By (1) and (2), and . It means that . This completes the sufficiency.
Now we prove the necessity. Suppose that . Then since . By the necessity of (1) and (2), G must be -free for each , . Therefore, G should be -free. By the definition of , . This completes the proof. □
By Theorems 1, 4 and 6, we can obtain the following corollary:
Corollary 1. Let S be a connected graph. Then G being a connected S-free graph implies if and only if S is an induced subgraph of .
Moreover, note that “G is -free” implies “” (see Theorems 4), this also means that if G is -free, then , here S can be any subgraph of G. Therefore, by Theorems 2, 5 and 6, we also can obtain the following corollary:
Corollary 2. Let be a set of two connected graphs such that R and S are not an induced subgraph of . Then G being a connected -free graph implies if and only if , , or .
3. The Necessity Part of Main Results
We first construct some families of connected graphs
(see
Figure 2. Here
). It is easy to see that each
satisfies
.
The necessity part of Theorem 4. Let S be a graph such that every connected S-free graph is . Then S is a common induced subgraph of all graphs in .
Note that the common induced subgraph of the graphs in and is a path. Since the largest induced path of the graphs in is , S must be an induced subgraph of . This completes the proof of the necessity part of Theorem 4.□
The necessity part of Theorem 5. Let R and S be not an induced subgraph of graphs such that every connected -free graph is . Then all graphs in should contain either R or S as an induced subgraph. Without loss of generality, we may assume that R is a common induced subgraph of all graphs in . Note that all graphs in contain no induced cycle with a length of at least 4 as an induced subgraph, so we need to consider the following four cases.
Case 1.R contains a clique with .
Since for , all graphs in are R-free, they all should contain S as an induced subgraph. Note that all graphs in are -free, and all graphs in are -free, so S should be a path. Since the largest induced path of the graphs in is , S should be an induced subgraph of , a contradiction.
Case 2.R does not contain the clique with , but contains two edge-disjoint .
Since R is a common induced subgraph of all graphs in , R should be . Since for , all graphs in are R-free, they all should contain S as an induced subgraph. Note that all graphs in are -free, and all graphs in are -free, so S should be a path. Since the largest induced path of the graphs in is , S should be an induced subgraph of . So .
Case 3.R does not contain the clique with , but contains exactly one .
Since R is a common induced subgraph of all graphs in , R should be an induced subgraph of . Since for , all graphs in are R-free, they all should contain S as an induced subgraph. Note that the common induced subgraph of all graphs in and is a tree with the maximum degree 3 or a path. If S is a tree with the maximum degree 3, since the common induced tree with the maximum degree 3 of all graphs in and are , S should be an induced subgraph of . So . If S is a path. Since the largest induced path of the graphs in is , S should be an induced subgraph of . So .
Case 4.R is a tree.
Since all graphs in are -free, R should be a path. Note that the largest induced path of the graphs in is , so R should be an induced subgraph of , a contradiction.
From the proofs above, we have that , , or . This completes the proof of the necessity part of Theorem 5.□
4. The Sufficiency Part of Main Results
In this section, we provide the sufficiency proof of main results.
The sufficiency part of Theorem 4. Let G be a connected -free graph. Then . If , G must be a complete graph and . If , by Theorem 3 (3), . This completes the proof of the sufficiency part of Theorem 4.□
The sufficiency part of Theorem 5. Let G be a connected -free graph such that , where , or . Then there must exist a minimum edge cut, say M, such that . Let and be the components of , and let , . Then , say , .
Claim 1. For , . Moreover, for any , .
Proof. We will count the number of edges of
for
.
Note that the complete graph has edges. It means that , i.e., .
Moreover, for any , since , . This completes the proof of Claim 1. □
Now we will distinguish the following two cases to complete our proof.
Case 1.G contains a with , and .
Subcase 1.1..
By Claim 1, there exist two vertices and such that . Then (if ), or (if ), or (if ), a contradiction.
Subcase 1.2..
By Claim 1, there exist two vertices and such that . Then (if ), or (if ), or (if ), a contradiction.
Subcase 1.3..
By Claim 1, and .
Suppose that or . Without loss of generality, we may assume that , it means there exist two vertices such that .
Then
, if ;
, if ;
, if ;
, if and ,
a contradiction.
Suppose that and . Note that and . Then and . Since and , and . Therefore and . It means that , , and . Since , the each vertex in is just adjacent to exactly one vertex which is in , and vice versa.
Suppose . Then there exists a vertex such that . Therefore
, if );
, if ;
, if ;
, if and ,
a contradiction.
Suppose . Then and .
Assume . Then there exists a vertex , such that and . Therefore (if ), or (if ), or (if ), a contradiction.
Assume . Then it means that and for any . Since and , there exist some vertices in such that their degree in G are at least 3. Then we choose a vertex , such that and as small as possible. Let be the shortest path between and . Then all inner vertices of should have degree two. Let and . Then contians an induced (if ), or contians an induced (if ), a contradiction.
Case 2.G contains no with , and .
Let , and for . Then and . By the minimality of M and the definition of , . Now we choose a path between and , such that and , and the length of the path is as small as possible. Then and all inner vertices of must be in . Let and , such that . Then is an induced path with at least 5 vertices, say .
Subcase 2.1..
is an induced path with at least 5 vertices, a contradiction.
Subcase 2.2..
By Claim 1, there exists a vertex such that . Then contians an induced (if ), or an induced (if ), a contradiction.
Subcase 2.3..
By Claim 1 and , there exist two vertices such that . Then contians an induced (if ), or an induced (if and ), or an induced (if or ), a contradiction.
This completes the proof of the sufficiency part of Theorem 5.□