摘要
介绍了一种利用量子行为粒子群算法(QPSO)求解多峰函数优化问题的方法。为此,在QPSO中引进一种物种形成策略,该方法根据群体微粒的相似度并行地分成子群体。每个子群体是围绕一个群体种子而建立的。对每个子群体通过QPSO算法进行最优搜索,从而保证每个峰值都有同等机会被找到,因此该方法具有良好的局部寻优特性。将基于物种形成的QPSO算法与粒子群算法(PSO)对多峰优化问题的结果进行比较。对几个重要的测试函数进行仿真实验结果证明,基于物种形成的QPSO算法可以尽可能多地找到峰值点,峰值收敛性能优于PSO。
An improved Quantum-behaved Particle Swarm Optimization （QPSO） for multi-peaks functions optimization was proposed, In the proposed Species-based QPSO （SQPSO）, the swarm population was divided into subpopulations in parallel based on their similarity. Each subpopulation was set around a dominating particle called the species seed. Over successive iterations, species were able to simultaneously optimize towards multiple optima by using QPSO, so each peaks were ensure to be searched equally, no matter whether they are global or local optima. Experimental results demonstrate that SQPSO can search as many peaks of function as possible. Simulation results show the convergence performance of SQPSO is superior to that of PSO.
出处
《计算机应用》
CSCD
北大核心
2006年第12期2956-2960,共5页
journal of Computer Applications
基金
国家自然科学基金资助项目(60474030)
关键词
量子行为粒子群算法
粒子群算法
物种形成策略
多峰寻优
Quantum-behaved Particle Swarm Optimization（ QPSO）
Particle Swarm Optimization（ PSO）
species
multi-peak searching
