摘要
针对微粒群算法在处理约束优化问题时,难以兼顾约束与优化之间关系的问题,提出了一种泛学习微粒群算法(ULPSO),通过引入微粒不可行历史最优,使得微粒的学习更具多样性和有效性,增强了算法的搜索智能。通过对常用的13个基准函数的测试对比分析,表明该算法求解约束优化问题的计算具有快速性、稳定性和有效性。
In dealing with constrained optimization problems, particle swarm optimization(PSO) is difficult to balance the relationship between constrains and optimization. A ubiquitous-learning swarm optimization (ULPSO) was proposed in order to solve this problem. Via incorporating infeasible personal best for each particle, this algorithm makes the study of particles more diversity and efficiency, enhances search intelligence. Finally, it is applied to a set of 13 well-known benchmark functions, and the experimental results illustrate its computing speed, stability and efficiency.
出处
《装备制造技术》
2009年第5期76-79,共4页
Equipment Manufacturing Technology
关键词
微粒群算法
约束优化
不可行历史最优
边界
学习
particle swarm optimization
constrained optimization
infeasible personal best
boundary
learn