资源描述：
《Fuzzy Adaptive Particle Swarm Optimization》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、FuzzyAdaptiveParticleSwarmOptimizationYuhuiShiEDSEmbeddedSystemsTeam1401E.HofferStreetKokomo,IN46902,USAYuhui.shi@eds.comandRussellC.EberhartDepartmentofElectricalandComputerEngineeringPurdueSchoolofEngineeringandTechnology799W.MichiganStreetIndianapolis,IN46202Eberhart@engr.iupui.eduAbs
2、tractparticlesaremanipulatedaccordingtothefollowingequation(ShiandEberhartI998a,I998b):Inthispaper,afuzzysystemisimplementedtodynamicallyadapttheinertiaweightoftheparticleswarmoptimizationalgorithm(PSO).Threebenchmarkfunctionswithasymmetricinitialrangesettingsareselectedasthetestfunction
3、s.ThesamefuzzysystemhasbeenappliedtoallthethreetestwherecIandc2arepositiveconstants,andrand()andfunctionswithdifferentdimensions.TheRand()aretworandomfunctionsintherange[O,I].XI=experimentalresultsillustratethatthefuzzyadaptive(xiI,xi?,...,xlD)representstheithparticle.Pi=(pll,p12,PSOisap
4、romisingoptimizationmethod,whichis...,P,~)representsthebestpreviousposition(theespeciallyusefulforoptimizationproblemswithapositiongivingthebestfitnessvalue)oftheithparticle.dynamicenvironment.Thesymbolgrepresentstheindexofthebestparticleamongalltheparticlesinthepopulation.V,=(v,,,v,~,1I
5、ntroduction...,vlD)representstherateofthepositionchange(velocity)forparticlei.Variablewistheinertiaweight.Particleswamoptimizationisapopulation-basedevolutionaryalgorithm.Itissimilartootherpopulation-Thebalancebetweenglobalandlocalsearchthroughoutbasedevolutionaryalgorithmsinthatthealgor
6、ithmisthecourseofruniscriticaltothesuccessofaninitializedwithapopulationofrandomsolutions.Itisevolutionaryalgorithm.Insomeimplementationsofunlikemostofotherpopulation-basedevolutionaryevolutionaryprogramming,balancebetweenglobalandalgorithms(Goldberg1989,Fogel1994,Rechenberglocalsearchis
7、obtainedthroughadaptingthevariance1994,Koza1992),however,inthatPSOismotivatedby(strategyparameter)oftheGaussianrandomfunctionorthesimulationofsocialbehaviorinsteadofsurvivalofstepsize.Furthermore,insomeimplementations,eventhefittest,andeachcandidatesolutionisassociatedwit
