A new view toward vertex decomposable graphs

https://doi.org/10.1016/j.disc.2022.112953Get rights and content

Abstract

In this paper, we bring a new view about closed neighbourhood to show the vertex decomposability of graphs. Making use of the characterization of hereditary vertex decomposable graphs, we introduce a class of vertex decomposable graphs, which include some well-known classic vertex decomposable graphs such as clique-whiskered graphs and Cameron-Walker graphs.

Introduction

As we know, vertex decomposable graphs are an important class of algebraic graphs, since vertex decomposable graphs are shellable, and well-covered shellable graphs are Cohen-Macaulay. It is known that chordal graphs form an important class of vertex decomposable graphs, this is characterized by R. Woodroofe in [14]. D. Cook and N. Nagel give a method to construct Cohen-Macaulay graphs in [2], actually, they construct a class of vertex decomposable graphs, called clique-whiskered graphs. In [6], T. Hibi et al. prove that Cameron-Walker graphs are vertex decomposable, and hence they are sequentially Cohen-Macaulay. Recently, Liu and Wu prove that Boolean graphs are well-covered and vertex decomposable in [8], which shows that Boolean graphs are a new class of Cohen-Macaulay graphs.

Recall that a simplicial complex (abbreviated as complex) Δ on a vertex set V is a finite subset of 2V, such that AΔ and BA implies BΔ. The set A is called a face if AΔ, and called a facet if A is a maximal face with respect to inclusion. The set of facets of Δ is denoted by F(Δ). The dimension of a face A, denoted by dim(A), is |A|1. The dimension of a complex Δ, denoted by dim(Δ), is the maximum dimension of its faces. The complex Δ is called pure if all the facets of Δ have the same dimension; otherwise, it is called nonpure. For a vertex xV, the link and the deletion of x in Δ is defined bylinkΔ(x)=F|xF,F{x}F(Δ),Δx=F|xFΔ.

A simplicial complex Δ is vertex decomposable if either Δ is a simplex (a complex with only one facet), or there exists a vertex x of Δ such that

(1) linkΔ(x) and Δx are vertex decomposable, and

(2) Each facet of linkΔ(x) is not a facet of Δx.

A vertex x satisfying conditions (1) and (2) is called a shedding vertex of Δ. If x only satisfies condition (2), then it is called a weak shedding vertex.

A simplicial complex Δ is called shellable if there exists a linear order F1,,Fm on its facet set F(Δ) such that for each pair i<j, there exists a k<j, such that |FjFk|=1 and FiFjFk. Correspondingly, if we require further that FkFiFj, then Δ is called strongly shellable. Note that on pure complexes, there are the following deductive relationships:

pure vertex decomposability ⟹ pure shellability ⟹ Cohen-Macaulayness.

For other concepts and results about simplicial complex, please refer to [5].

Let G be a simple graph. Denote by V(G) and E(G) its vertex set and edge set respectively. A subset T of V(G) is called an independent set if each pair of vertices in T are nonadjacent in G. An independent set is called maximal if it is maximal with respect to inclusion. A graph is well-covered if all maximal independent sets have the same cardinality.

A simplicial complex Δ is called the independence complex of a graph G, denoted by Δ(G), if the facet set F(Δ) consists of all maximal independent sets of the graph G. It is clear that Δ(G) is pure if and only if G is well-covered. For a graph G, it is said to be vertex decomposable (shellable, respectively), if its independence complex Δ(G) is vertex decomposable (shellable, respectively). Now assume further G has vertex set V(G)={x1,,xn} with edge set E(G), and let R=K[x1,x2,,xn] be the polynomial ring with n variables x1,,xn over a field K. Recall that the edge ideal associated to G is defined to be the ideal I(G)=xixj|{xi,xj}E(G) of R, and it is well known that I(G) is identical with the Stanley-Reisner ideal IΔ of the simplicial complex Δ=Δ(G). The edge ideal of a graph was introduced and the Cohen-Macaulay property of the edge ideal was studied by Villarreal, see [13] for details.

This paper is organized as follows: In chapter 2, we introduce a view to see vertex decomposability of a graph by closed neighbourhoods. Then, hereditary vertex decomposable graphs are characterized in chapter 3. Making use of the view about closed neighbourhood, a class of vertex decomposable graphs which are called core vertex decomposable graphs are introduced in chapter 4. Finally, in chapter 5, a class of well-covered vertex decomposable graphs, which are actually clique-whiskered graphs, are restudied.

Section snippets

Vertex decomposability characterized by closed neighbourhoods

Let G be a simple graph. For each vertex a in G, its neighbourhood in G, denoted by N(a), consists of all vertices which are adjacent to a in G, i.e., N(a)={b|abE(G)}. The closed neighbourhood of a in G is defined by N[a]=N(a){a}. In order to distinguish neighbourhoods from closed neighbourhoods, in this paper, neighbourhoods are called open neighbourhoods. Sometimes we use NG(a) (NG[a], respectively) to denote the open (closed, respectively) neighbourhood of a in G to emphasize the basic

Hereditary vertex decomposable graphs

Hereditary shellable simplicial complexes were studied in [4]. In this paper, we focus on the study on graphs. Recall that a simple graph G is hereditary shellable, if every induced subgraph of G is shellable. Similarly, one can define hereditary vertex decomposable graphs and hereditary sequentially Cohen-Macaulay graphs.

Definition 3.1

Let G be a simple graph. If every induced subgraph of G is vertex decomposable (sequentially Cohen-Macaulay, respectively), then G is called a hereditary vertex decomposable

A class of vertex decomposable graphs

Definition 4.1

Let G be a simple graph. If the subgraph of G induced by M[G] is hereditary vertex decomposable, then G is called a core vertex decomposable graph.

Theorem 4.2

If G is a core vertex decomposable graph, then G is vertex decomposable.

Proof

We will prove the result by induction on |V(G)|. If |V(G)|=1, then the result is clear. Assume that the result holds for |V(G)|<n. In the following, consider the case while |V(G)|=n. If M[G]=V(G), then G is vertex decomposable since the subgraph induced by M[G] is vertex

A class of well-covered vertex decomposable graphs

Proposition 5.1

Let G be a simple graph. If G is well-covered, then the subgraph of G induced by M[G] is also well-covered.

Proof

Let G be the subgraph of G induced by M[G]. If G is not well-covered, then we can find two maximal independent sets A and B in G, with |A|<|B|. Since G is well-covered, so A is not a maximal independent set in G. Hence there exists aV(G)A, such that A{a} is an independent set. If aM[G], then A{a} is an independent set in G, which contradicts with the fact that A is a maximal

Declaration of Competing Interest

This paper does not have any conflicts to disclose.

References (14)

There are more references available in the full text version of this article.

This research was supported by the National Natural Science Foundation of China (Grant No. 11961017, 11971338), the Natural Science Foundation of Shanghai (No. 19ZR1424100), the Hainan Provincial Natural Science Foundation of China (Grant No. 119MS002), and the Young Talents' Science and Technology Innovation Project of Hainan Association for Science and Technology (Grant No. QCXM201806).

View full text