粒子群算法的收敛性和多样性分析与优化

Analyze and Optimize the Convergence and Diversity of Particle Swarm Optimization

针对粒子群优化算法的收敛性和多样性问题,提出一种基于混沌优化的震荡粒子群优化算法。该算法利用混沌特性和震荡环节扩大粒子搜索的遍历性,将混沌状态引入到优化变量使粒子获得持续搜索的能力。实验测试证明该算法不仅能保持种群的多样性,而且能有效避免算法陷入早熟收敛现象。

According to the convergence and diversity of particle swarm optimization, this paper proposes a chaos concussion particle swarm optimization algorithm （CCPSOA） which uses the properties of chaos and concussion link to lead particles＇ exploration, so as to equip the swarm system with an ability of sustainable search. Numerical simulation demonstrates that not only can the proposed algorithm keep the population diversity, but also ef- fectively prevent the premature convergence phenomenon.