A conjecture on the eigenvalues of threshold graphs

From MaRDI portal



Abstract: Let An be the anti-regular graph of order n. It was conjectured that among all threshold graphs on n vertices, An has the smallest positive eigenvalue and the largest eigenvalue less than −1. Recently, in cite{Cesar2} was given partial results for this conjecture and identified the critical cases where a more refined method is needed. In this paper, we deal with these cases and confirm that conjecture holds.


A simple graph \(G = (V, E)\) is called a threshold graph if there exists a function \(w: V \rightarrow [0,\infty)\) and a non-negative real number \(t\) such that \(uv \in E\) if and only if \(w(u) + w(v) \ge t\). A simple graph is called an anti-regular graph if only two vertices of it have equal degrees. It is known that, up to isomorphism, there is exactly one connected anti-regular graph on \(n\) vertices, which is denoted as \(A_n\). Anti-regular graphs form a subfamily of the family of threshold graphs. In [\textit{C. O. Aguilar} et al., Linear Algebra Appl. 557, 84--104 (2018; Zbl 1396.05064)], it was conjectured that for each \(n\), \(A_n\) has the smallest positive eigenvalue and the largest negative eigenvalue less than \(-1\) among all threshold graphs on \(n\) vertices. In [\textit{C. O. Aguilar} et al., ibid. 588, 210--223 (2020; Zbl 1437.05128)], this conjecture was proved for all threshold graphs on \(n\) vertices except for \(n - 2\) critical cases where the interlacing method fails. In this paper, the author deals with these cases and thereby completes the proof of this conjecture.











This page was built for publication: A conjecture on the eigenvalues of threshold graphs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2228534)