摘要
粒子群算法(PSO)中参数的选择是一个重要研究方向,参数的设置常依靠经验来确定,从而造成工作量大且难以得到最优的参数组合,影响了算法的使用。针对以上情况,本文使用3个测试函数对粒子群算法和收缩因子方法(CFM)中的收缩因子、速度约束和种群规模等重要参数进行了系统的实验和分析,并且提出了参数取值策略。实验证明本文提出的参数取值策略能明显地改进PSO算法性能,具有一定的实用价值。
The selection of the parameters in particle swarm optimization(PSO) is an important research field.In general,the parameters are determined by experience and experiment.This leads to heavy work load and difficulties to obtain the optimal combination of the parameters,hence,affecting the use of PSO.Aiming at this condition,the effects of the major parameters such as the constriction factor,velocity constraint and population size in PSO and constriction factor method(CFM) are systematically investigated based on three benchmark functions in the papper.The authors present a recommended setting strategy of parameters on CFM and PSO,which can remarkably improve the performance of the PSO algorithm,and the setting strategy is of practical value.
出处
《西华大学学报（自然科学版）》
CAS
2008年第1期76-80,共5页
Journal of Xihua University:Natural Science Edition
关键词
粒子群算法
参数选择
进化计算
particle swarm optimization(PSO)
parameter selection
evolutionary computation
