The average connectivity matrix of a graph

TL;DR

This paper explores spectral properties of the average connectivity matrix of a graph, establishing bounds related to maximum matchings and graph structure, with specific results for bipartite graphs.

Contribution

It introduces bounds on the spectral radius of the average connectivity matrix, linking it to maximum matchings and providing characterizations for equality cases.

Findings

Spectral radius of the average connectivity matrix is bounded by graph parameters.

For any connected graph, the spectral radius is at most 4 times the maximum matching size divided by n.

For bipartite graphs, a tighter bound on the spectral radius is established, with equality for complete balanced bipartite graphs.

Abstract

For a graph $G$ and for two distinct vertices $u$ and $v$, let $\kappa(u,v)$ be the maximum number of vertex-disjoint paths joining $u$ and $v$ in $G$. The average connectivity matrix of an $n$-vertex connected graph $G$, written $A_{\bar{\kappa}}(G)$, is an $n\times n$ matrix whose $(u,v)$-entry is $\kappa(u,v)/{n \choose 2}$ and let $\rho(A_{\bar{\kappa}}(G))$ be the spectral radius of $A_{\bar{\kappa}}(G)$. In this paper, we investigate some spectral properties of the matrix. In particular, we prove that for any $n$-vertex connected graph $G$, we have $\rho(A_{\bar{\kappa}}(G)) \le \frac{4\alpha'(G)}n$, which implies a result of Kim and O \cite{KO} stating that for any connected graph $G$, we have $\bar{\kappa}(G) \le 2 \alpha'(G)$, where $\bar{\kappa}(G)=\sum_{u,v \in V(G)}\frac{\kappa(u,v)}{{n\choose 2}}$ and $\alpha'(G)$ is the maximum size of a matching in $G$; equality holds only when $G$ is a complete graph with an odd number of vertices. Also, for bipartite graphs, we improve the bound, namely $\rho(A_{\bar{\kappa}}(G)) \le \frac{(n-\alpha'(G))(4\alpha'(G) - 2)}{n(n-1)}$, and equality in the bound holds only when $G$ is a complete balanced bipartite graph.