Table of Contents
Fetching ...

Characterizing the positive inertia index of connected signed graphs in terms of girth

Suliman Khan, Sakander Hayat, Mohammed J. F. Alenazi

TL;DR

This work bounds the positive inertia index $p^{+}(G^{\sigma})$ of a connected signed graph by its girth $g_r$ and derives sharp extremal classifications. By extending interlacing techniques and cycle-signature analysis from ordinary graphs to signed graphs, the authors prove $p^{+}(G^{\sigma}) \geq \left\lceil \frac{g_r}{2} \right\rceil - 1$ and characterize when equality occurs, revealing two primary extremal families: balanced cycles with specific parity and balanced complete multipartite graphs. The results further sharpen to classify all graphs with $p^{+}(G^{\sigma}) = \left\lceil \frac{g_r}{2} \right\rceil$ and provide a detailed taxonomy of canonical unicyclic and certain bicyclic signed graphs achieving this bound. Overall, the paper extends Duan and Yang's 2024 findings for ordinary graphs to the signed-graph setting, enriching the relationship between inertia indices and girth in spectral graph theory.

Abstract

Let $G^σ=(G,σ)$ be a connected signed graph and $A(G^σ)$ be its adjacency matrix. The positive inertia index of $G^σ$, denoted by $p^{+}(G^σ)$, is defined as the number of positive eigenvalues of $A(G^σ)$. Assume that $G^σ$ contains at least one cycle, and let $g_{r}$ be its girth. In this paper, we prove $p^{+}(G^σ) \geq \lceil \frac {g_{r}}{2} \rceil-1$ for a signed graph $G^σ$. The extremal signed graphs corresponding to $p^{+}(G^σ) = \lceil \frac {g_{r}}{2} \rceil-1$ and $p^{+}(G^σ) =\lceil \frac {g_{r}}{2} \rceil$ are characterized, respectively. The results presented in this article extend the recent work on ordinary graphs by Duan and Yang (Linear Algebra Appl., 2024) to the context of signed graphs.

Characterizing the positive inertia index of connected signed graphs in terms of girth

TL;DR

This work bounds the positive inertia index of a connected signed graph by its girth and derives sharp extremal classifications. By extending interlacing techniques and cycle-signature analysis from ordinary graphs to signed graphs, the authors prove and characterize when equality occurs, revealing two primary extremal families: balanced cycles with specific parity and balanced complete multipartite graphs. The results further sharpen to classify all graphs with and provide a detailed taxonomy of canonical unicyclic and certain bicyclic signed graphs achieving this bound. Overall, the paper extends Duan and Yang's 2024 findings for ordinary graphs to the signed-graph setting, enriching the relationship between inertia indices and girth in spectral graph theory.

Abstract

Let be a connected signed graph and be its adjacency matrix. The positive inertia index of , denoted by , is defined as the number of positive eigenvalues of . Assume that contains at least one cycle, and let be its girth. In this paper, we prove for a signed graph . The extremal signed graphs corresponding to and are characterized, respectively. The results presented in this article extend the recent work on ordinary graphs by Duan and Yang (Linear Algebra Appl., 2024) to the context of signed graphs.

Paper Structure

This paper contains 3 sections, 17 theorems, 23 equations, 2 figures.

Key Result

Theorem 1

Let $G^\sigma=(G,\sigma)$ be a connected signed graph with girth $g_{r}$. Then $p^{+}(G^\sigma) \geq \lceil \frac{g_{r}}{2} \rceil-1$, where the equality holds if and only if: (i) $G^\sigma=C^\sigma_{g_{r}}$ is a balanced signed cycle satisfying $g_{r} \equiv 0,1 \pmod{4}$ or an unbalanced signed cy

Figures (2)

  • Figure 1: A canonical unicyclic graph with $p^+(G^\sigma) = \lceil \frac{6}{2} \rceil = 3$ having positive eigenvalues $1.41421,~1.73205$, and $2.44949$.
  • Figure 2: Bicyclic signed graphs $B^\sigma(5, 3, 5)$, $B^\sigma(5, 5, 5)$ and $B^\sigma(5, 4, 5)$, where the doted lines represent the negatived edges.

Theorems & Definitions (23)

  • Theorem 1
  • Theorem 2
  • Theorem 3
  • Theorem 4
  • Lemma 1
  • Lemma 2
  • Lemma 3
  • Lemma 4
  • Lemma 5
  • Corollary 1
  • ...and 13 more