Table of Contents
Fetching ...

On the maximum spectral radius of connected graphs with a prescribed order and size

Ivan Damnjanović

TL;DR

The paper resolves the problem of maximizing the spectral radius $\rho(G)$ over connected graphs with order $n$ and size $n-1+e$ by proving that extremality occurs within threshold graphs and is captured by the templates $\mathcal{D}_{n,e}$ and $\mathcal{V}_{n,e}$. A computer-assisted framework based on $\mathcal{T}$-subgraphs reduces the search to a finite set of templates, enabling complete verification for $e\le 130$, together with explicit comparison criteria via a cubic root threshold $\psi_e$ and a Bell-type argument for large $e$ that establishes $\mathcal{V}_{n,e}$ as the unique maximizer when $n\ge e+2+13\sqrt{e}$. The results yield a precise, regime-based extremal characterization and provide computable mechanisms (via polynomials $\mathcal{P}_{\mathcal{T}_1(G)}$, $\mathcal{P}_{\mathcal{T}(G)}$, and $\mathcal{Q}_{G_1,G_2}$) to decide which graph attains the maximum spectral radius. Overall, the work advances extremal spectral graph theory by delivering complete solutions for $e\le 130$ and sharp large-$n$ bounds, incorporating both combinatorial and algebraic methods alongside computer-assisted verification.

Abstract

The spectral radius of a graph is the largest modulus of an eigenvalue of its adjacency matrix. Let $\mathcal{C}_{n, e}$ be the set of all the connected simple graphs with $n$ vertices and $n - 1 + e$ edges. Here, we solve the spectral radius maximization problem on $\mathcal{C}_{n, e}$ when $e \le 130$ or $n \ge e + 2 + 13\sqrt{e}$.

On the maximum spectral radius of connected graphs with a prescribed order and size

TL;DR

The paper resolves the problem of maximizing the spectral radius over connected graphs with order and size by proving that extremality occurs within threshold graphs and is captured by the templates and . A computer-assisted framework based on -subgraphs reduces the search to a finite set of templates, enabling complete verification for , together with explicit comparison criteria via a cubic root threshold and a Bell-type argument for large that establishes as the unique maximizer when . The results yield a precise, regime-based extremal characterization and provide computable mechanisms (via polynomials , , and ) to decide which graph attains the maximum spectral radius. Overall, the work advances extremal spectral graph theory by delivering complete solutions for and sharp large- bounds, incorporating both combinatorial and algebraic methods alongside computer-assisted verification.

Abstract

The spectral radius of a graph is the largest modulus of an eigenvalue of its adjacency matrix. Let be the set of all the connected simple graphs with vertices and edges. Here, we solve the spectral radius maximization problem on when or .

Paper Structure

This paper contains 5 sections, 29 theorems, 92 equations.

Key Result

Theorem 1.1

Suppose that $G$ attains the maximum spectral radius on $\mathcal{G}_m$ for some $m \in \mathbb{N}_0$. Then $G$ is a threshold graph.

Theorems & Definitions (43)

  • Theorem 1.1: bruhoff1985
  • Theorem 1.2: bruhoff1985
  • Theorem 1.3: rowlinson1988
  • Theorem 1.4: brusol1986
  • Theorem 1.5: brusol1986
  • Theorem 1.6: brusol1986
  • Theorem 1.7: cvetrow1988
  • Theorem 1.8: bell1991
  • Theorem 1.9: oleroydri2002
  • Proposition 1.10
  • ...and 33 more