An Improved Competitive Many-Objective Particle Swarm Optimization Algorithm Based on Two-Round Selection Method

Particle swarm optimization (PSO) algorithms have been proposed for solving many-objective optimization problems (MaOPs). However, the tradeoff between convergence and diversity is still a challenge. In this paper, a two-round environmental selection strategy is proposed to pursue a good tradeoff between diversity and convergence for many-objective PSO (MaPSO). The two-round selection strategy is embedded into a hypervolume indicator-based competitive MaPSO framework for forming the algorithm CMaPSO-HT. We compare it with five state-of-the-art algorithms to further demonstrate the performance of CMaPSO-HT on benchmark MaOPs. The experimental results show that CMaPSO-HT outperforms these existing algorithms in terms of convergence and population diversity.