自适应搜索区域的微粒群优化算法  

Particle swarm optimization algorithm based on self-adaptive search area

作　　者：裴振奎[1] 韩锦峰[1] 李华[1] 宋建伟[1] 

机构地区：[1]中国石油大学(华东)计算机与通信工程学院,山东东营257061

出　　处：《计算机工程与设计》2008年第14期3732-3734,共3页Computer Engineering and Design

摘　　要：基于基本微粒群优化算法搜索后期,众多微粒都拥挤在历史最优位置周围进行重复性无效搜索这一现象,提出一种改进的微粒群算法——自适应搜索区域的微粒群优化算法,其主要思想为:每当搜索进行到当前设定的一个最大迭代次数时(即,微粒在全局历史最优位置周围徘徊进行无效搜索时),在原搜索区域的基础上,重新构造一个较小的搜索区域,并重新初始化微粒,继续进行搜索,最终获得最优解。对3个常用标准测试函数进行优化计算,仿真结果表明,该算法具有比基本微粒群优化算法更好的优化性能。Based on the phenomena that a lot of particles crowded around the best position and many particles repeated an ineffective search in search later period, an improved algorithm is proposed, which is called particle swarm optimization algorithm based on self-adaptive searcharea （SSAPSO）. Its characteristic was that while alot ofparticles crowded around the best position and repeated an ineffective search, there constructed a new less search area based on former one, initialized particles renewedly and continued search, The global optimization solution is obtained in the end. Then, both SSAPSO and PSO are used to resolve three widely used test functions＇ optimization problems. The experimental results indicate that SSAPSO has better optimization performance than PSO.

关 键 词：微粒群优化算法 自适应 搜索区域 优化 微粒 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]