Extremal eigenvalues with respect to graph minors

TL;DR

This paper investigates the maximum spectral radius of graphs excluding certain minors, revealing structural properties and extremal values, and extends known results to new classes of minors using advanced spectral and structural techniques.

Contribution

It introduces a unified framework for extremal spectral radius in minor-free graphs, including new structural characterizations and extremal values for various minors.

Findings

Graphs with spectral radius above a threshold contain specific minors or spanning books.

Characterization of extremal graphs as minor-saturated after removing dominating vertices.

Extension of known results to new minors like flowers, wheels, and generalized books.

Abstract

Let $spex(n,H_{minor})$ denote the maximum spectral radius of $n$-vertex $H$-minor free graphs. The problem on determining this extremal value can be dated back to the early 1990s. Up to now, it has been solved for $n$ sufficiently large and some special minors, such as $\{K_{2,3},K_4\}$, $\{K_{3,3},K_5\}$, $K_r$ and $K_{s,t}$. In this paper, we find some unified phenomena on general minors. Every graph $G$ on $n$ vertices with spectral radius $\rho\geq spex(n,H_{minor})$ contains either an $H$ minor or a spanning book $K_{\gamma_H}\nabla(n-\gamma_H)K_1$, where $\gamma_H=|H|-\alpha(H)-1$. Furthermore, assume that $G$ is $H$-minor free and $\Gamma^*_s(H)$ is the family of $s$-vertex irreducible induced subgraphs of $H$, then $G$ minus its $\gamma_H$ dominating vertices is $\Gamma^*_{\alpha(H)+1}(H)$-minor saturate, and it is further edge-maximal if $\Gamma^*_{\alpha(H)+1}(H)$ is a connected family. As applications, we obtain some known results on minors mentioned above. We also determine the extremal values for some other minors, such as flowers, wheels, generalized books and complete multi-partite graphs. Our results extend some conjectures on planar graphs, outer-planar graphs and $K_{s,t}$-minor free graphs. To obtain the results, we combine stability method, spectral techniques and structural analyses. Especially, we give an exploration of using absorbing method in spectral extremal problems.