Table of Contents
Fetching ...

The Antisymmetric Line Graph

Hartosh Singh Bal

Abstract

We introduce the \emph{antisymmetric line graph} $\mathcal{A}(G)$, a canonical signed refinement of the line graph defined on the oriented edges of a graph $G$ via an antisymmetric incidence rule. We show that $\mathcal{A}(G)$ is balanced if and only if $G$ is bipartite, so the frustration index $\ell(\mathcal{A}(G))$ defines a switching-invariant measure of non-bipartiteness. The switching class of $\mathcal{A}(G)$ determines $G$ up to isomorphism modulo isolated vertices, thereby resolving Whitney's exceptional ambiguity. We obtain quantitative bounds relating $\ell(\mathcal{A}(G))$ to classical measures of non-bipartiteness; in particular, \[ \operatorname{def}(G)\le \ell(\mathcal{A}(G))\le (Δ(G)-1)\operatorname{def}(G), \] where $\operatorname{def}(G)$ is the minimum number of edges whose deletion makes $G$ bipartite. Our strongest result is an exact identification in the cubic case: for every cubic graph $G$, with $\operatorname{oct}(G)$ denoting the odd cycle transversal number, \[ \ell(\mathcal{A}(G)) = 2\operatorname{oct}(G). \] Hence on cubic graphs a canonical signed line-graph invariant recovers a central bipartization parameter, and computing $\ell(\mathcal{A}(G))$ is NP-hard even for cubic inputs.

The Antisymmetric Line Graph

Abstract

We introduce the \emph{antisymmetric line graph} , a canonical signed refinement of the line graph defined on the oriented edges of a graph via an antisymmetric incidence rule. We show that is balanced if and only if is bipartite, so the frustration index defines a switching-invariant measure of non-bipartiteness. The switching class of determines up to isomorphism modulo isolated vertices, thereby resolving Whitney's exceptional ambiguity. We obtain quantitative bounds relating to classical measures of non-bipartiteness; in particular, where is the minimum number of edges whose deletion makes bipartite. Our strongest result is an exact identification in the cubic case: for every cubic graph , with denoting the odd cycle transversal number, Hence on cubic graphs a canonical signed line-graph invariant recovers a central bipartization parameter, and computing is NP-hard even for cubic inputs.
Paper Structure (35 sections, 20 theorems, 56 equations)

This paper contains 35 sections, 20 theorems, 56 equations.

Key Result

Theorem 1

Let $G$ and $H$ be connected graphs. If $L(G)\cong L(H)$, then either $G\cong H$, or $\{G,H\}=\{K_3,K_{1,3}\}$.

Theorems & Definitions (43)

  • Theorem 1: Whitney Whitney1932
  • Definition 1: Switching class
  • Lemma 1: ALG refines the line graph
  • proof
  • Definition 2: Antisymmetric Line Graph
  • Remark 1: Laplacian spectrum versus switching class
  • Proposition 1: Switching invariance
  • proof
  • Proposition 2: Canonical signing
  • proof
  • ...and 33 more