摘要
比例因子F的合适赋值常会大大改善差分进化算法的求解性能,但是如何给值是个麻烦的事情.本文给出了二种免比例因子F的差分进化算法.算法将每一个个体视为带电粒子,利用之间的吸引、排斥机制,确定个体在差分方向上移动的长度,依此免去比例因子F设置的麻烦.通过和两种PSO算法以及其它四种不同赋值策略的算法的数值试验比较,表明提出的算法相比其它相比较的算法有更好的求解性能.
A fit setting of the scale factor F can usually improve greatly the performance of diferenfial evolution,however, how to set is nuisance. Two diffcrential evolutions without scale factor F are presented in the paper. The algorithms look upon each individuals as a charged particle and utilize the attraction-repulsion mechanism of the particles to decide on the step length of the motion of the individual in the direction of the difference for the purpose of avoiding the setting of the scale factor F. The comparisons of numerical experiments among the proposed algorithms,two PSO algorithms and four other algorithms with the different setring strategies are done, which show that the performance of the proposed algorithms outperform other compared algorithms.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2009年第6期1318-1323,共6页
Acta Electronica Sinica
基金
国家自然科学资金资助(No.60674108,60705004)
综合业务网理论及关键技术国家重点试验室基金资助(No.ISN02080003)
关键词
类电磁机制
全局优化
粒子群优化
差分进化
electromagnetism-like mechanism
global optimization
particle swarm optimization
differential evolutionl
