基于两阶段策略的粒子群优化

被引:6
作者
徐俊杰
忻展红
机构
[1] 北京邮电大学经济管理学院
关键词
粒子群优化; 优化算法; 全局优化; 计算智能; 群体智能;
D O I
暂无
中图分类号
TP301.6 [算法理论];
学科分类号
081202 ;
摘要
提出了一种基于传统粒子群优化的两阶段实施方案,通过对一组测试函数的仿真表明,该方案以适当增加计算量为代价,提高了搜索成功率.对比实验表明,两阶段方案几乎在各种最大可迭代次数的约束下都能获得更好的搜索成功率,且对学习速度参数的敏感性降低,算法的搜索性能更稳健.实施该策略时,子群数量宜选取适中的数值,以综合考虑可靠性与计算成本2个因素.
引用
收藏
页码:136 / 139
页数:4
相关论文
共 8 条
[1]  
Hybrid particleswarm optimiser with breeding and subpopulation. Lovbjerg M,Rasmussen T K,Krink T. Proceedings of the Genetic and Evolutionary ComputationConference . 2001
[2]  
Comparing inertia weights and con-striction factors in particle swarm optimization. Eberhart R,Shi Y. Pro-ceedings of the 2000 Congress on Evolutionary Computa-tion . 2000
[3]  
Particle swarm optimization. Kennedy J,Eberhart R. Proceedings of IEEE International Conference on Neu-ral Networks . 1995
[4]  
An off-the-shelf PSO. Carlisle A,Dozier G. http:∥antho.huntingdon.edu/publications/Off-The-Shelf-PSO.pdf . 2005
[5]  
The swarm and the queen:towards a determin-istic and adaptive particle swarm optimization. Clerc M. Pro-ceedings of the 1999 Congress of Evolutionary Computa-tion . 1999
[6]  
Guest editorial:special issue onparticle swarm optimization. Eberhart R C,Shi Y. IEEE Trans on Evolu-tionary Computation . 2004
[7]  
A new optimizer using parti-cle swarm theory. Eberhart R C,Kennedy J. Proceedings of the Sixth Interna-tional Symposium on Micro Machine and Human Science . 1995
[8]  
A modified particle swarm opti-mizer. Shi Y,Eberhart R C. Proceedings of the IEEE International Con-ference on Evolutionary Computation . 1998