Table of Contents
Fetching ...

On the Number of Vertices/Edges whose Deletion Preserves the Konig-Egervary Property

Vadim E. Levit, Eugen Mandrescu

TL;DR

This work investigates how deleting vertices or edges preserves the König-Egerváry property, focusing on graphs where $\alpha(G)+\mu(G)=n(G)$. By leveraging critical independent sets and the associated core/ker structure, the authors derive a precise vertex heredity formula $\varrho_v(G)=n(G)-\xi(G)+\varepsilon(G)$ for König-Egerváry graphs, and establish a tight, generally stronger-than-baseline edge heredity bound $\varrho_e(G)\ge m(G)-\xi(G)+\varepsilon(G)$. The analysis hinges on the behavior of induced subgraphs on critical sets, the decomposition $G=G[X]\cup G-X$ with $X=A\cup N(A)$, and the extendability of maximum matchings, yielding deep links between deletions and maximum matching structure. These results advance understanding of how near-König-Egerváry properties behave under vertex/edge deletions and connect to Lorentzen-type inequalities via the critical-set framework.

Abstract

The graph G=(V,E) is called Konig-Egervary if the sum of its independence number and its matching number equals its order. Let RV(G) denote the number of vertices v such that G-v is Konig-Egervary, and let RE(G) denote the number of edges e such that G-e is Konig-Egervary. Clearly, RV(G) = |V| and RE(G) = |E| for bipartite graphs. Unlike the bipartiteness, the property of being a Konig-Egervary graph is not hereditary. In this paper, we present an equality expressing RV(G) in terms of some graph parameters, and a tight inequality bounding RE(G) in terms of the same parameters, when G is Konig-Egervary.

On the Number of Vertices/Edges whose Deletion Preserves the Konig-Egervary Property

TL;DR

This work investigates how deleting vertices or edges preserves the König-Egerváry property, focusing on graphs where . By leveraging critical independent sets and the associated core/ker structure, the authors derive a precise vertex heredity formula for König-Egerváry graphs, and establish a tight, generally stronger-than-baseline edge heredity bound . The analysis hinges on the behavior of induced subgraphs on critical sets, the decomposition with , and the extendability of maximum matchings, yielding deep links between deletions and maximum matching structure. These results advance understanding of how near-König-Egerváry properties behave under vertex/edge deletions and connect to Lorentzen-type inequalities via the critical-set framework.

Abstract

The graph G=(V,E) is called Konig-Egervary if the sum of its independence number and its matching number equals its order. Let RV(G) denote the number of vertices v such that G-v is Konig-Egervary, and let RE(G) denote the number of edges e such that G-e is Konig-Egervary. Clearly, RV(G) = |V| and RE(G) = |E| for bipartite graphs. Unlike the bipartiteness, the property of being a Konig-Egervary graph is not hereditary. In this paper, we present an equality expressing RV(G) in terms of some graph parameters, and a tight inequality bounding RE(G) in terms of the same parameters, when G is Konig-Egervary.
Paper Structure (5 sections, 45 theorems, 42 equations, 12 figures)

This paper contains 5 sections, 45 theorems, 42 equations, 12 figures.

Key Result

Lemma 1.1

LevMan2022 If $G$ is an almost bipartite graph, then $n(G)-1\leq\alpha(G)+\mu(G)\leq n(G)$.

Figures (12)

  • Figure 1: $G_{1}-v_{1}$, $G_{2}-e_{2}$ are König-Egerváry graphs, while $G_{1}-v_{2}$ and $G_{2}-e_{1}$ are not König-Egerváry graphs.
  • Figure 2: $S\in\Omega(G)$ and $A$ is a critical independent set of $G$.
  • Figure 3: $G$ is a non-König-Egerváry graph with core$(G)=\{a,b,c,u\}$.
  • Figure 4: $d\left( G_{1}\right) =d\left( G_{2}\right) =1$, while only $G_{1}$ is not a König-Egerváry graph.
  • Figure 5: $G_{1}$, $G_{2}$ and $G_{3}$ are König-Egerváry graphs
  • ...and 7 more figures

Theorems & Definitions (46)

  • Lemma 1.1
  • Theorem 1.2
  • Theorem 1.3
  • Theorem 1.4
  • Theorem 1.5
  • Theorem 1.6
  • Corollary 1.7
  • Proposition 1.8
  • Corollary 1.9
  • Theorem 2.1
  • ...and 36 more