Co-operative Vector-Evaluated Particle Swarm Optimization for Multi-objective Optimization

Abstract

Vector-evaluated particle swarm optimization is a particle swarm optimization variant which employs multiple swarms to solve multi-objective optimization problems. Recently, three variants of particle swarm optimization which utilize co-operative principles were shown to improve performance in single-objective environments. This work proposes co-operative vector-evaluated particle swarm optimization algorithms, which employ co-operative particle swarm optimization variants within vector-evaluated particle swarm optimization swarms. Performance of the proposed algorithms is compared with the standard vector-evaluated particle swarm optimization algorithm using various knowledge transfer strategies. A comparison of the best performing co-operative vector-evaluated particle swarm optimization variants is also made against well-known multi-objective PSO algorithms. Each co-operative vector-evaluated particle swarm optimization variant significantly outperforms standard vector-evaluated particle swarm optimization with respect to the hyper volume metric, with two of three variants also yielding improved solution distribution. The results indicate that co-operation is a powerful tool which enhances hyper volume and solution distribution of the original vector-evaluated particle swarm optimization algorithm, allowing co-operative vector-evaluated particle swarm optimization variants to successfully compete with top multi-objective PSO optimization algorithms.