A Distribution Evolutionary Algorithm for Graph Coloring

Abstract

Graph Coloring Problem (GCP) is a classic combinatorial optimization problem that has a wide application in theoretical research and engineering. To address complicated GCPs efficiently, a distribution evolutionary algorithm based on population of probability models (DEA-PPM) is proposed. Based on a novel representation of probability model, DEA-PPM employs a Gaussian orthogonal search strategy to explore the probability space, by which global exploration can be realized using a small population. With assistance of local exploitation on a small solution population, DEA-PPM strikes a good balance between exploration and exploitation. Numerical results demonstrate that DEA-PPM performs well on selected complicated GCPs, which contributes to its competitiveness to the state-of-the-art metaheuristics.