Mean Square Convergent Particle Swarm Optimization Based on Fuzzy Migratory Operator

This paper proposes a novel cooperative particle swarm PSO (particle swarm optimization)algorithm, which makes the use of the property of the fuzzy migratory operator to achieve the optimization performance. To avoid the drawback of the possibility of being trapped in local optimum, the proposed method uses a migrate-based strategy to control the diversity of the swarm. During the iteration aspect of the algorithm, the comprehensive fuzzy evaluation method is employed to evaluate the diversity. Furthermore, the fuzzy migratory operator is then used to remove bad particles once the diversity is far from ideal. Moreover, we have proven that the proposed migrate strategy is a mean square convergence. The experimental results conducted on three benchmark functions also proved that the proposed method is superior to that of classical PSO and conventional cooperative PSO, where the comparison has been based primarily upon the global optimality, solution accuracy and diversity value.