This article foucuses on -free graph. In this paper, we prove that if G is -free, then . We then use this result to obtain the upper bound of order and chromatic number of -free graph .
A graph consists of its vertex set and edge set . The order of , denoted by , is the size of . The complement of a graph , denoted by or , is a graph with the same vertex set while whose edge set consists of the edges not present in .
All graphs in this paper are finite and have no loops or parallel edges. A graph contains if is isomorphic to an induced subgraph of . If is isomorphic to , then we say . For a collection of graphs , is -free if it does not contain . In this paper, if does not contain , we will say that is edge-free.
Path and cycle on vertex is denoted by and , respectively. The complete graph on that has vertex is denoted bu . We use to denote the disjoint union of and .
A -coloring of a graph is a function such that whenever and are adjacent in . We say that is -colorable if admits a -coloring. The chromatic number of is denoted by which represents the minimum positive integer such that is -colorable. The clique number is denoted by which represents the size of the largest clique in .
A family of graphs is said to be -bounded if there exists a function f such that for every graph and every induced subgraph of it holds that . The function is called a -binding function for . The class of perfect graphs (a graph is perfect if for every induced subgraph of it holds that , for instance, is a -bounded family with -binding function . Therefore, -boundedness is a generalization of perfection. The notion of -bounded families was introduced by Gyárfás who make the following conjecture.
conjecture 1.1.
[3]
For every forest T, the class of -free graphs is -bounded.
For , Esperet, Lemoine, Maffray and Morel[2] obtained the optimal bound on the chromatic number: every -free graph is 5-colorable. Serge Gaspers and Shenwei Huang[6] obtained the optimal bound of the chromatic number: every -free graph is 4-colorable. Karthick and S. Mishra obtained the optimal bound of the chromatic number: every -free graph is 6-colorable. Rui Li,Jinfeng Li and Di Wu claimed that [5] -free graph is -colorable. We follow the methods in this paper and sharper the bound to .
Partition into two following parts:
•
•
Let be a -hole. We use to denote a graph obtained from by connecting edges and .
Let be a -hole. We use to denote a graph obtained from by connecting edges , and .
Our proof includes three major theorem:
theorem 1.1.
If contains , then .
theorem 1.2.
If contains or , then .
theorem 1.3.
If is -free and is not an induced subgraph of , then .
Finally, we obtain a theorem for -free graphs:
theorem 1.4.
If is -free with clique number and order n, then and .
2 Structure Around
Suppose there is a in and its vertex set is . Naturally, we can divide
into following three sets:
•
.
•
.
•
.
we need to combine and and divide it into three disjoint sets: . The set includes vertices that is only adjacent to and vertices in . The set includes vertices that are only adjacent to . Since is anticomplete to an edge in , is -free.
can be partitioned in to (mod3) for . It is easy to obtain that (mod3) for .
3 main theorem
We first introduce a lemma provided by Rui Li, Jinfeng Li and Di Wu.
Proof. If there is an edge in (or ), then (or ), which contradicts definition of (or ).∎
We fisrt introduce a lemma to help us slightly determine the structure of .
theorem.
(1.1)
If contains , then .
Proof. Suppose , . we can divide into . Since is -free, is unions of clique and . Therefore, .
We will prove and hence . can be divided into , , , and . We apply claim 3.3 to prove that the clique number of first four sets are at most 1. If the last set has edge, then . Therefore, . ∎
observation 1.
If is -free, then is -free.
Now we introduce a lemma which is important in the proof of theorem1.2.
claim 3.3.
Suppose can be partitioned into three nonempty subsets , and such that and are -free graph, is a stable set , is -free , is -free graph for and any vertex in , has no edge. Furthermore, for any vertex , if , then . If has at most one edge, then .
Proof.
If has no edge, then randomly select one vertex . We partition accoding to , that is . . Now we partition into three stable set: , and . is stable set as is -free. According to definition, the rest two sets are stable sets.
If has one edge, then set one vertex of the vertex set of that edge . We partition into three part: . It is obvious that , otherwise induces an or . Therefore, as is -free, we can conclude that . and hence . ∎
observation 2.
If contains , then according to section 2.1, can be partitioned into and .
lemma 3.4.
If contains , then .
Proof.
. Otherwise there exists an edge in . We call the edge . If , then or (depending on whether is adjacent to ). As a consequence, both and are not adjacent to . However, is complete to , which can be proven through making use of -free. Similarly, is complete to . As a consequence, , which contradicts the fact that . As a consequence, .
For any , must be one of four following sets: ,, and . We partition () into (),(),() and () accoding to their neigbors in . Trivially , for .
If , then there is no edge in . Suppose and , we are going to prove are complete to and anticomplete to and hence , which contradicts . If is adjacent to , then is isomorphic to or . Symmetrically, is anticomplete to . As consequence, must be complete to , or will induce . Symmetrically, is complete to . As consequence, we get the contradiction that . Therefore, .
If , then there is no edge in , which is similar to the proof for
From dicussion above, if , then and hence .
Suppose and . Furthermore, . Otherwise, induces or . According to the definition, can be partitioned into , and .
Suppose there is a vertex in has only one neighbor in . It is not difficult to obtain that the only neighbor is . If we exchange and in , then we can prove in the same way as what we did when .
Suppose evrey vertex in has two neighbors in .
Let be a and be two vertex outside . We use co-donino1 to denote a family of graphs obtained from by connecting edges and can be connected or disconnected.
Let be a and be two vertex outside . We use co-domino2 to denote a graph obtained from by connecting edges and .
Suppose . The following case shows that if is complete to or complete to , then . In other words, if contains co-domino1 or co-domino2, then .
Recalling that is anticomplete to , is anticomplete to .
case1: contains or .
Suppose contains .
contains at most one edge. Otherwise, suppose contains more than one edge. Since is -free and complete to and anticomplete to , there exists an such that these edges are isomorphic to and has at most one neighbor in every edges or induces . However, leads to a contradiction that induces .
contains at most one edge. Since is -free and complete to and anticomplete to , this union of edges is isomorphic to for some and has at most one neighbor in every edge or induces . However, leads to a contradiction that induces .
Let , then according to claim3.3, . Furthermore, .
Let Let be a and be two vertex outside . We use co-domino3 to denote a graph obtained from by connecting edges and .
Suppose and . The following case shows that if exists, then . In other words, if contains , then .
Partition into following five vertex sets.
.
.
.
case2: Suppose contains .
Suppose .
. Define . Recalling that is anticomplete to and that is anticomplete to , is anticomplete to . Therefore, is anticomplete to . Since , is a stable set and hence . Similarly, . Trivially, is stable set and hence the chromatic number is less than 2. Therefore,
is a stable set. If , then contains . Accoding to definition, . The rest is obvious.
is a stable set. Recalling that and , . According to definition is complete to and the rest is obvious.
is a stable set. According to definition, and is a stable set. The rest is obvious.
Let be .
Suppose .
is a stable set. If , then contains . Since , .
.
Combining -free, -free and our supposition that evrey vertex in has two neighbors in , if , then is complete to . Similarly, if ,then is complete to . Therefore, is a stable set.
Furthermore,. Noticing that , the proof is easy.
Therefore, if ,
Suppose . We continue to use notations . Similar to case2, .(The proof does not depend on the existence on {x}.)∎
Let be a -hole and be a vertex outside the -hole. We use to denote a graph obtained from by connecting edges and .
Let be a -hole and be a vertex outside the -hole. We use to denote a graph obtained from by connecting edges , and .
The following figures are and ( from left to right).
lemma 3.5.
Suppose is -free and contains or , then .
Proof.
case1:G contains .
is anticomplete to and complete to .Suppose and (or ) only, then . Therefore, must complete to both and and hence we get ,which contradics our assumption. is complete to (or ) or will induce an ( will induce a ).
Symmetrically , is anticomplete to and complete to and is anticomplete to and anticomplete to .
Either or is edge-free. According to the above paragraph, is complete to and hence it is -free. Suppose there is an edge in () and an edge in ().
We prove to obtain a contradicion. It is easy to see that should have at least one neighbor in . So does . Consequently, , otherwise must contain .
Symmetrically, either or is edge-free and either or is edge-free.
. Recalling that is -free. Suppose has edge, then . If or contains edge, then we can prove . If , and are all edge-free, then it is trvial that .
Therefore,
case2: is -free and contains .
Accoring to section 2.1, we partition around .
Every vertex in is adjacent to and not adjacent to . If there is such that is not adjacent to , then is adjacent to . Otherwise or . Therefore, we can simply suppose there is such that is adjacent to . However, or (depending on the adjacency of and ). Therefore, such does not exist.
Every vertex in is adjacent to .
Every vertex in is anticomplete to .
Partition into the folloing six sets.
•
.
•
.
•
.
•
•
•
Since is compete to and , has no edge and hence .
Since and anticomplete to , is -free and hence .
. Since is anticomplete to and , has no edge. Since is anticomplete and (It is trivial that is anticomplete to ), has no edge.
Since is anticomplete to and , has no edge and hence has no edge.
. Therefore, ∎
Suppose all graphs are -free.
Let be a -hole. We use to denote a graph obtained from by adding a vertex which is only adjacent .
Let be a -hole and be a vertex outside the -hole. We use to denote a family of graphs obtained from by connecting edge and and it does not matter whether is connected or not.
Define . In this lemma, does not include and does not include . We introduce a partition which is useful in the following lemma.
observation 3.
If contains . can be partitioned into and .
lemma 3.6.
If contains , then .
Proof.
case1: G contains an element of .
is complete to and anticomplete to . Otherwise, suppose there is which is not adjacent to or has neighbor in . If has neighbor in , then induces or . Therefore, is not adjacent to and hence is anticomplete to . Therefore, such does not exist.
Similarly, is complete to and anticomplete to .
Either or has no edge. Otherwise, suppose there is an edge() in and an edge() in . If induces or , then induces or . Therefore, is not isomorphic to , and . However, such condition requires to be isomorphic to or . Therefore, there must be a vertex has degree one in . Suppose in . Then is anticomplete to and hence . Therefore, either does not exist or does not exists.
is complete to . Otherwise, suppose there exists such that z is not adjacent to . is adjacent to , otherwise . Since , is adjacent to . However, if is adjacent to , then . Therefore, such does not exist.
contains at most one edge. Suppose there are two edges in ( and ). Because is complete to . has at most one neighbor in and . Suppose . Since both are not adjacent to , , which causes contradiction. Therefore, has at most one edge.
Suppose has no edge. Let , then we can apply claim3.3 to get and hence .
Noticing that vertices in have neighbor in , we can divide to make the structure more clearly. We difine two disjoint sets as following:
•
is adjacent to only one vertex in
•
is adjacent to more than one vertices in
. Otherwise, suppose and . Suppose is adjacent to .
Because if the only neighbor of belongs to , then induces or .
If is not adjacent to , then is isomorphic to an element in (when is not adjacent to . However, if is not adjacent to , then . Therefore, such does not exist.
Any vertex in is complete to an edge in . Otherwise, there is which is not complete to any edge in . must be adjacent to , which leads to the contradiction that induces or .
can be colored with no more than colors.
Recalling that points in are at least adjacent to one edge in , it is trivial that . In summary, . ∎
Suppose all graphs are -free.
Let be a -hole. We use to denote a graph obtained from by connecting edge .
lemma 3.7.
If contains or -, then .
Proof.
case1: contains .
We prove is edge-free. We obtain contradiction by proving . In other words, contradiction comes out if are anticomplete to and complete to . Suppose there is an edge in and we call them . If (or ), would induce or . Furthermore, we can prove is complete to without much difficulty and hence we finish our proof of proposion that is edge-free.
Furthermore, either or has no edge. Otherwise, suppose there is , . Since is complete to and anticomplete to and is complete to and anticomplete to , induces or .
is complete to and hence is edge-free. Otherwise, there is such that has only one neighbor in . If , then . Otherwise, induces . However, induces . Therefore, . Now , otherwise induces . However, induces . Therefore, such does not exist.
every edge in includes one vertex which is complete to . Otherwise, there is an edge in whose vertices has at most one neighbor in . If one of them() is adjacent to rather than ,then induces . Therefore, none of these vertices is adjacent to and hence induces . We already obtain a contradiction.
According to the above discussion, can be partitioned into two stable sets: the vertices which are complete to and the rest of vertices in . Therefore, .
Therefore, ∎
Combining lemma3.4 and lemma3.7, we obtain theorem1.2 straightly.
theorem.
1.2
If contains or , then .
Suppose is -free with is -free.
claim 3.8.
If and , then either or has no edge.
Proof.
Suppose both sets have edges. The vertex sets in is and that of edge in is . must induce or or . Therefore, either or is edge-free.
∎
theorem.
(1.3)
If is not a clique, then 7.
Proof.
Since is not a clique, has two nonadjacent vertices . According to claim3.8, either or has no edge. Without loss of generality, we suppose is edge-free and hence .
If , then or . We suppose , then . If has edge(), then accoring to section 2.1, can be divided into .
. The vertex sets left to be colored are: and . Since , and . Therefore,
Suppose .
We partition into three parts: , and . Define .
. Suppose there is an edge() in . Then can be divided into and and . Since is anticomplete to and , must be -free and hence . Similarly, . Because and , . Therefore ,.
. According to definition of , there is a induced in . Define and . Since is --free, is anticomplete to (mod). Obviously, for and is edge-free. Therefore .
Because , . Consequently, =7. ∎
Therefore, combining theorem1.1, theorem1.2 and theorem1.3, we obtain our major theorem.
theorem.
If is -free, then .
Finally, we prove a simple theorem using the above theorem.
theorem.
(1.4)
If is -free with order and clique number , then and .
Proof.
Suppose is -free.
If is connected, then we apply the above theorem to obtain that . Therefore, can be partitioned into stable sets and hence can be partitioned into cliques. Therefore, .
If is not connnected, then there is such that and hence . Such decomposition will last until each satidfied is connected. Suppose . Since and , .
It is trivial that the complement of bipartite graph is perfect graph. Therefore can be partitioned into perfect graphs and one clique. Therefore, .∎
4 acknowledge
I try to go further on -free graphs. However, I fail and 7 is the best upper bound I obtain. I know this is not a well-written paper. If you are confused about some details in the proof, you can email me at [email protected] or [email protected].
References
[1]A. Bharathi,S. Choudum. Colouring of -free graphs.Graphs and Combinatorics 34(1):97–107,2018.
[2]L. Esperet, L. Lemoine, F. Maffray, and G. Morel. The chromatic number of -free
[3]A. Gyárfás. On ramsey covering numbers. Coll. Math. Soc. János Bolyai, 10:801–816, 1973.
graphs. Discrete Math, 313:743–754, 2013.
[4]T. Karthick and S. Mishra. Coloring (, diamond, )-free graphs.arXiv:1703.00606, 2017.
[5]Rui Li, Jinfeng Li and Di Wu. On the chromatic number of some ()-free graphs, arXiv: 2308.15248, 2023.
[6]G. Serge and S. Huang. -Free Graphs are 4-Colorable. SIAM Joural on Discrete Mathematics 33:1095–1120,2019.