摘要
分析了基本粒子群算法(PSO)全局搜索能力与收敛速度的矛盾,提出了粒子群相似度的概念.根据每个粒子与全局最优粒子的不同相似度,对基本PSO算法的惯性权重进行动态调整.同时提出一种根据相似度计算聚集度的方法,并根据聚集度的大小随机地对粒子重新赋值,控制粒子群的多样性,提高了全局搜索能力.典型优化问题的实例仿真验证了该算法的有效性.
The contradiction of the global exploration and convergence speed of particle swarm optimization（PSO） algorithm is analyzed. A new PSO algorithm is proposed, in which the inertia weight of every particle will be changed dynamically with the similarity between the particle and the current optimal position.And based on similarity, a method of collection calculation is proposed, which is used to randomly initialize the position of the particle, control the diversification of PSO, and improve the ability of global exploration. Experiments on benchmark functions show the effectiveness of the new algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2007年第10期1155-1159,共5页
Control and Decision
关键词
粒子群算法
全局最优性
相似度
聚集度
Particle swarm optimization algorithm
Global optimality
Similarity
Collection
