首页 | 本学科首页   官方微博 | 高级检索  
     

改进的变尺度混沌粒子群算法及其应用
引用本文:王俭臣,单甘霖,刘东平,赵志宁.改进的变尺度混沌粒子群算法及其应用[J].计算机工程与设计,2012,33(5):1973-1977,2012.
作者姓名:王俭臣  单甘霖  刘东平  赵志宁
作者单位:1. 军械工程学院光学与电子工程系,河北石家庄,050003
2. 解放军63898部队,河南济源,459000
3. 军械工程学院导弹工程系,河北石家庄,050003
基金项目:国防预研基金项目(513270203)
摘    要:针对粒子群算法(PSO)存在局部最优及后期收敛速度慢等问题,提出一种改进的变尺度混沌粒子群算法(IMCPSO).该算法初期,在整个解空间对最优粒子进行变尺度混沌扰动,以防止陷入局部最优;算法后期,则以最优粒子为中心引入变尺度混沌扰动,以提高算法收敛速度.当算法一旦陷入局部最优时,采用混沌粒子替代部分种群粒子以增加粒子多样性,使算法尽快跳出局部最优.基于benchmark测试函数的仿真结果表明,所提算法与基本粒子群算法(SPSO)和变尺度混沌粒子群算法(MCPSO)相比,具有明显好的搜索精度和收敛速度.最后,将该算法应用于电路故障诊断实验中的支持向量机参数优化问题,实验结果说明了其应用价值.

关 键 词:粒子群算法  混沌  变尺度  支持向量机  参数优化  故障诊断

Improved mutative scale chaos particle swarm optimization and its application
WANG Jian-chen , SHAN Gan-lin , LIU Dong-ping , ZHAO Zhi-ning.Improved mutative scale chaos particle swarm optimization and its application[J].Computer Engineering and Design,2012,33(5):1973-1977,2012.
Authors:WANG Jian-chen  SHAN Gan-lin  LIU Dong-ping  ZHAO Zhi-ning
Affiliation:1.Department of Optics and Electronics Engineering,Ordnance Engineering College,Shijiazhuang 050003,China; 2.Unit 63898 of PLA,Jiyuan 459000,China;3.Department of Missile Engineering,Ordnance Engineering College,Shijiazhuang 050003,China)
Abstract:To overcome the shortcomings of local minima and low convergence speed in particle swarm optimization(PSO),an improved mutative scale chaos particle swarm optimization(IMCPSO) is proposed.At the early phase of this optimization,mutative scale chaotic disturbance on the whole solution space is applied to avoid local convergence.And at the late phase,the optimal particle centered mutative chaotic disturbance,which could improve the convergence character,is used.In case of local convergence,to jump out of it,using chaotic particles to replace some swarm particles to increase particle variety.Simulation results on benchmark test functions indicated that the proposed optimization has rather better searching precision and convergence speed than basic PSO(SPSO) and mutative scale chaos PSO(MCPSO).Finally,the proposed algorithm is applied to SVM parameter optimization problem in a circuit fault diagnosis experiment and the experimental results show the application value of it.
Keywords:particle swarm optimization  chaos  mutative scale  support vector machine  parameter optimization  fault diagnosis
本文献已被 CNKI 万方数据 等数据库收录!
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