An improved particle swarm optimization for multi-objective discrete optimization

Abstract

In this paper an improved multi-objective particle swarm optimization algorithm (IMOPSO) is designed to efficiently solve multi-objective discrete optimization problems. In the IMOPSO, a novel similarity-based selecting scheme is used to selection of the global best solution and individual best solution for each particle, and an external set truncation strategy is used to maintain the diversity in the Pareto optimal solutions. Additionally, a local search subroutine is applied on every particle to improve the search efficiency of optimization. The IMOPSO is compared with two multi-objective particle swarm optimization algorithms proposed in the literature on several test problems, and experimental results show that the IMOPSO has good performance in multi-objective discrete optimization