摘要
针对粒子群可能会陷入局部极优值所导致的早熟问题,提出一种有限作用域的混沌粒子群优化方法。利用特定的初始分布涵盖全局最优值,利用混沌序列良好的非线性性质,影响粒子速度更新过程,增加粒子种群的多样性。以有限作用域外的粒子遍历优化问题的可行域,增加粒子对可行域的广度搜索,以作用域内的粒子搜索最优值,提高全局最优值的精度搜索效率。数值实验表明提出的算法优于标准粒子群和传统的混沌粒子群,并能解决粒子群的早熟问题。
To solve the premature problem that the swarm is trapped by local optimization in searching process,the Limited Effective Region Chaotic Particle Swarm Optimization（LER-CPSO） is proposed.It applies the special initialization to cover the global optimization.It increases the swarm diversities by introducing the chaotic series nonlinear property in the velocity update process.The particles outside the limited effective region explore the feasible region to improve the exploration.The particles in the limited effective region search the optimization to improve the precision.The numerical experiment shows that the proposed algorithm has better searching results than the standard PSO and the typical CPSO.It solves the premature problem.
出处
《计算机工程与应用》
CSCD
北大核心
2011年第12期39-42,共4页
Computer Engineering and Applications
关键词
Logistic序列
混沌粒子群优化
有限作用域
群智能
Logistic series
Chaotic Particle Swarm Optimization（CPSO）
limited effective region
swarm intelligence
作者简介
庞淑萍（1972－），女，副教授，主要研究领域为智能优化算法、系统工程与规划。E-mail：meatbal11982@163．com
