Abstract

Recently, more and more studies to apply particle swarm optimization (PSO) algorithm for solving multi-objective problems (MOPs) have been proposed. Due to there will be more than one global best solution gbest found can be solution searching clue in MOPs. For leading more particles toward to potential searching space and find global optimum, suitable local guide (global best solution) assignment becomes an important issue in multi-objective particle swarm optimizer. This paper presents a density-based local guide assign approach called balanced guide assignment (BGA) for multi-objective particle swarm optimization. Besides, for making proposed method more robust, a cluster archive approach and perturbation method are also involved with BGA. They can efficient improve particles' solution searching ability to find more solutions located on/near to the Pareto front. Six benchmarks were adopted for testing and compare the proposed method with other related works. From the results, the proposed method performed better in performance metrics can be observed.

Paper Details