摘要
针对粒子群算法在求解复杂的多峰问题时极易陷入局部最优解的问题,提出一种基于K-均值聚类的动态多种群粒子群算法(KDMSPSO).在该算法中,利用K-均值聚类算法将种群分成若干个子群(聚类);为了增强子群间的信息交流,对子群进行动态重组;在每个子群中,粒子的速度由它所在子群的中心粒子和该粒子所有邻居的信息共同调整.在基准函数测试和实际应用中,其结果显示KDMSPSO算法相比其他PSO算法具有一定的优势.
Particle swarm optimizer（PSO）may easily get trapped in a local optimum,when it comes to solving complex multimodal problems.Therefore,this paper presents dynamic multi-swarm particle swarm optimizer based on K-means clustering（KDMSPSO）.In KDMSPSO,the population is divided into several sub-swarms by using K-means clustering.In order to increase the message exchange of sub-swarms,the sub-swarm is dynamically constructed,and the velocity of each particle is adjusted by clustering center that it belongs to and all particles in its neighborhood including itself.In benchmark function and actual application,the experimental results show that the KDMSPSO algorithm can achieve better solutions than other PSO algorithms.
出处
《控制与决策》
EI
CSCD
北大核心
2011年第7期1019-1025,共7页
Control and Decision
基金
贵州教育厅社科项目(0705204)
遵义师范学院基础教育课题(基07017
基07015)
遵义市科技局项目([2008]21)
关键词
粒子群算法
K-均值
动态多种群
particle swarm optimizer
K-means
dynamic multi-swarm
