首页 | 本学科首页   官方微博 | 高级检索  
     

基于趋利避害原则的粒子群算法研究
引用本文:王星博,李本威,杨欣毅,贾明明.基于趋利避害原则的粒子群算法研究[J].计算机应用研究,2012,29(3):933-936.
作者姓名:王星博  李本威  杨欣毅  贾明明
作者单位:1. 海军航空工程学院飞行器工程系,山东烟台,264001
2. 海军航空工程学院青岛分院,山东青岛,266041
摘    要:针对标准粒子群算法进行多极点函数优化时易导致早熟收敛及陷入局部最优的问题,把生物学中昆虫生存的趋利避害原则引入到粒子群优化算法中,改变传统粒子群优化算法只存在趋利操作而没有避害操作的单向性,提出了两种不同的保持或增加种群多样性的改进算法。仿真实验结果表明,与传统粒子群优化算法相比,采用基于趋利避害原则的粒子群算法处理复杂的多峰函数可显著提高算法的全局寻优性能。

关 键 词:粒子群优化  趋利避害  种群多样性  标准测试函数

Research on particle swarm optimizing algorithm based on seek advantage and avoid disadvantage principle
WANG Xing-bo,LI Ben-wei,YANG Xin-yi,JIA Ming-ming.Research on particle swarm optimizing algorithm based on seek advantage and avoid disadvantage principle[J].Application Research of Computers,2012,29(3):933-936.
Authors:WANG Xing-bo  LI Ben-wei  YANG Xin-yi  JIA Ming-ming
Affiliation:(1. Dept. of Airborne Vehicle Engineering, Naval Aeronautical Engineering Academy, Yantai Shandong 264001, China; 2.Qingdao Branch, Naval Aeronautical Engineering Academy, Qingdao Shandong 266041, China)
Abstract:To get rid of resulting in premature convergence and plunging into local optimum for standard particle swarm optimization to solve multiple-order pole functions,this paper introduced seek advantage and avoid disadvantage principle of some insects in biology into particle swarm optimization algorithm to change the single direction characteristic of the traditional algorithm which only had drawing on advantage operation instead of avoiding disadvantage operation.It proposed both kinds of ameliorated algorithms to maintain or increase population diversity.Simulation experiment results indicate that the particle swarm optimization based on seek advantage and avoid disadvantage principle can prominently improve the global optimization ability of the algorithm when dealing with complicated multimodal functions.
Keywords:particle swarm optimization  seek advantage and avoid disadvantage  population diversity  benchmark function
本文献已被 CNKI 万方数据 等数据库收录!
点击此处可从《计算机应用研究》浏览原始摘要信息
点击此处可从《计算机应用研究》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