基于静态有向种群结构的EPSO算法研究  

Research of Extended Particle-swarm Optimization Algorithm of Static Directed Population Structure

作　　者：台亚丽 曾建潮[1] 莫思敏[1] 

机构地区：[1]太原科技大学复杂系统与计算智能实验室,太原030024

出　　处：《太原科技大学学报》2014年第1期38-43,共6页Journal of Taiyuan University of Science and Technology

基　　金：山西省青年科技研究基金(2012021012-5);太原科技大学校博士科研启动基金(20122055)

摘　　要：较弱的局部搜索能力是影响扩展的微粒群算法(EPSO)性能的重要因素。为了提高扩展微粒群算法的局部搜索能力,本文针对EPSO的特点设计了各种静态有向种群结构。通过研究静态有向种群结构的特征度量与EPSO算法性能的关系,得出微粒的出度值和适应值是影响EPSO局部搜索与全局搜索能力的重要因素等结论。The weak local searching capability is an important factor influencing the performance of Extended Parti- cle Swarm Optimization （EPSO）algorithm. In order to improve the local search capability of EPSO, various static directed structures were inves according to the characteristics of EPSO. Some conclusions such as the impor- tant factor influence of outdegree of nodes and the particles fitness values on the global and the local search capa- bility of EPSO, were drawn in simulation experiment by studying the relations among characteristics measures of static directed structures and performances of EPSO.

关 键 词：扩展微粒群算法(EPSO) 静态有向结构 搜索能力 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]