摘要： 针对粒子群优化算法易陷入局部最优的问题,提出一种动态分组的粒子群优化算法。通过对鸟群习性的研究,给出交互粒子的概念,并在粒子群优化过程中引入动态分组机制,将种群动态划分成多个子种群,且每次划分的子种群数目是从特定集合中随机选取,从而增加交互粒子划分到同一子种群的概率。每个子种群在收敛进化的同时,利用环拓扑结构提高种群多样性及算法搜索全局最优解的能力。实验结果表明,与其他粒子群优化算法相比,该算法具有更好的稳定性、寻优性能以及更高的收敛精度。

关键词: 粒子群优化, 局部最优, 全局最优, 交互粒子, 动态分组, 环拓扑结构

Abstract: Aiming at Particle Swarm Optimization(PSO) algorithm is easy to fall into local optimal problems,this paper puts forward a PSO algorithm of dynamic group.Through the study of the flock behavior,the concept of interacting particles is presented.It introduces dynamic groupings into the PSO algorithm.Population is divided into multiple sub populations dynamically,and the number of each division of sub populations is randomly selected from a specific set.It increases the probability of interacting particles into the same sub population.During converging evolution,each sub population uses the ring topology structure to increase the diversity of population and the global search ability of the algorithm.Experimental results show that compared with other PSO algorithms,the algorithm has better optimal performance,stability,and higher convergence precision.

Key words: Particle Swarm Optimization(PSO), local optimum, global optimum, interacting particle, dynamic grouping, ring topology

中图分类号: 

• TP18