Bounding Clique-Width via Perfect Graphs

TL;DR

This paper investigates the clique-width of graphs that exclude certain subgraphs, identifying new classes with bounded or unbounded clique-width, and explores the complexity implications for graph isomorphism and coloring problems.

Contribution

It introduces three new classes of graphs with bounded clique-width and one with unbounded clique-width, using novel techniques involving perfect graphs and clique covering bounds.

Findings

Three classes have bounded clique-width

One class has unbounded clique-width

Graph Isomorphism is complete on the unbounded class

Abstract

Given two graphs $H_1$ and $H_2$, a graph $G$ is $(H_1,H_2)$-free if it contains no subgraph isomorphic to $H_1$ or $H_2$. We continue a recent study into the clique-width of $(H_1,H_2)$-free graphs and present three new classes of $(H_1,H_2)$-free graphs of bounded clique-width and one of unbounded clique-width. The four new graph classes have in common that one of their two forbidden induced subgraphs is the diamond (the graph obtained from a clique on four vertices by deleting one edge). To prove boundedness of clique-width for the first three cases we develop a technique based on bounding clique covering number in combination with reduction to subclasses of perfect graphs. We extend our proof of unboundedness for the fourth case to show that Graph Isomorphism is Graph Isomorphism-complete on the same graph class. We also show the implications of our results for the computational complexity of the Colouring problem restricted to $(H_1,H_2)$-free graphs.