论文标题
在横向临界图上定理的光谱加强
Spectral strengthening of a theorem on transversal critical graphs
论文作者
论文摘要
图$ g $的横向组是一组向所有边缘的顶点事件。 $ g $的横向数,用$τ(g)$表示,是$ g $的横向组的最低基数。如果$τ(g-e)<τ(g)$对于e(g)$中的每个边缘$ e \,则无隔离顶点的简单图$ g $被称为$τ$ - 至关重要。对于任何$τ$ - 关键图$ g $,带有$τ(g)= t $,已证明$ | v(g)| \ le 2t $ byerdős和gallai,以及$ | e(g)| \ le {t+1 \ select 2} $ reerdős,hajnal and hajnal and Moon和Moon。最近,Gyárfás和Lehel将其扩展到$ | v(g)| + | e(g)| \ le {t+ 2 \选择2} $。在本文中,我们通过频谱证明了更强的结果。令$ g $为$τ$ - 批评$τ(g)= t $和$ | v(g)| = n $,让$λ_1$表示$ g $的邻接矩阵的最大特征值。我们表明,当$ n+λ_1\ le 2t+1 $具有等值时,并且仅当$ g $是$ tk_2 $,$ k_ {s+1} \ cup(t-s)k_2 $,或$ c_ {2s-1} \ cup(t s)特别是,当$λ_1(g)\ le t $具有平等时,仅当$ g $为$ k_ {t+1} $时。然后,我们将其应用于表明,对于任何非负整数$ r $,我们都有$ n \ left(r+\ frac {λ_1} {2} {2} {2} \ right)\ le {t+r+1 \ select 2} $并表征所有极端图。这意味着$ r | v(g)| + | e(g)| \ le {t+r+1 \选择2} $,它比erdős-hajnal-moon定理和gyárfás-lehel定理强。我们还有其他一些概括。
A transversal set of a graph $G$ is a set of vertices incident to all edges of $G$. The transversal number of $G$, denoted by $τ(G)$, is the minimum cardinality of a transversal set of $G$. A simple graph $G$ with no isolated vertex is called $τ$-critical if $τ(G-e) < τ(G)$ for every edge $e\in E(G)$. For any $τ$-critical graph $G$ with $τ(G)=t$, it has been shown that $|V(G)|\le 2t$ by Erdős and Gallai and that $|E(G)|\le {t+1\choose 2}$ by Erdős, Hajnal and Moon. Most recently, it was extended by Gyárfás and Lehel to $|V(G)| + |E(G)|\le {t+2\choose 2}$. In this paper, we prove stronger results via spectrum. Let $G$ be a $τ$-critical graph with $τ(G)=t$ and $|V(G)|=n$, and let $λ_1$ denote the largest eigenvalue of the adjacency matrix of $G$. We show that $n + λ_1\le 2t+1$ with equality if and only if $G$ is $tK_2$, $K_{s+1}\cup (t-s)K_2$, or $C_{2s-1}\cup (t-s)K_2$, where $2\leq s\leq t$; and in particular, $λ_1(G)\le t$ with equality if and only if $G$ is $K_{t+1}$. We then apply it to show that for any nonnegative integer $r$, we have $n\left(r+ \frac{λ_1}{2}\right) \le {t+r+1\choose 2}$ and characterize all extremal graphs. This implies a pure combinatorial result that $r|V(G)| + |E(G)| \le {t+r+1\choose 2}$, which is stronger than Erdős-Hajnal-Moon Theorem and Gyárfás-Lehel Theorem. We also have some other generalizations.