多策略粒子群算法及应用

In this paper in order to solve the problem that the canonical particle swarm algorithm is prone to fall into the local optima and low convergence accuracy, a multi strategy improved particle swarm optimization algorithm (MSIPSO) is proposed. The population is divided into two sub groups: dominant subgroup and common subgroup based on fitness of particle. Both sub groups use targeted evolutionary strategies. The particles in dominant subgroup are mainly responsible for save the excellent information and global exploration, while particles in common subgroup use the excellent information of the particles in the dominant subgroup to improve themselves and to improve the quality of the whole population solution. In order to verify the performance of the proposed algorithm, MSIPSO compares with other PSO-based competitors andMulti-strategy particle swarm optimization and application Difference algorithm . It gets the best results on different benchmark functions, and it is proved that the proposed algorithm has better performance. In addition, the proposed algorithm is used to solve the portfolio problem, it also obtained the better results.