首页 | 本学科首页   官方微博 | 高级检索  
     

均匀粒子群算法
引用本文:刘宏达,马忠丽.均匀粒子群算法[J].智能系统学报,2010,5(4):336-341.
作者姓名:刘宏达  马忠丽
作者单位:哈尔滨工程大学 自动化学院,黑龙江 哈尔滨 150001
基金项目:中央高校基本科研业务专项资金资助项目 
摘    要:由于粒子群算法(PSO)本质上的随机性,其搜索质量和速度也呈随机性.这使得普通的粒子群算法难以满足某些需要快速优化的工程需要.利用均匀设计方法产生PSO算法的初始种群(或关键代次种群),可以使种群中的粒子在搜索空间分布更均匀,更好地保持分散性.算法中给出了4种种群的生成方案,通过测试和对比分析表明:基于值域分割的均匀设计种群生成法能使算法的搜索效果最好;算法可以在不丧失搜索精度和效率的前提下,提高搜索效率和搜索精度的稳定性,有效减少粒子聚集和早熟的发生.

关 键 词:粒子群  均匀设计  实时优化  关键代次种群

A particle swarm optimization algorithm based on uniform design
LIU Hong-da,MA Zhong-li.A particle swarm optimization algorithm based on uniform design[J].CAAL Transactions on Intelligent Systems,2010,5(4):336-341.
Authors:LIU Hong-da  MA Zhong-li
Affiliation:College of Automation, Harbin Engineering University, Harbin 150001,China
Abstract:In a normal PSO algorithm, particle populations are usually produced randomly, leading to variations in search quality and speed. Such algorithms cannot be used to solve engineering problems that must be optimized quickly. In order to solve such problems, a PSO algorithm based on uniform design was used to generate initial PSO populations. This made the distribution of particles more uniform in the search space. Four methods for generation of particle swarms were studied. Test results showed that search precision was considerably improved if the generation of a particle swarm by the uniform design method was based on value range division. That method maintained search efficiency and precision, stabilized search efficiency and precision, and reduced the possibility of swarm aggregation and premature convergence. 
Keywords:PSO  uniform design  real time optimization  key population
本文献已被 万方数据 等数据库收录!
点击此处可从《智能系统学报》浏览原始摘要信息
点击此处可从《智能系统学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号