粒子群算法及其改进技术研究

31()Vol312JournalofHubeiNormalUniversity(NaturalScience)No2,2011李志,陈年生,郭小珊,柯宗武(湖北师范学院计算机科学与技术学院,湖北黄石435002):粒子群算法是一种基于种群的随机优化技术,1995年由Eberhart博士和Kennedy博士提出,该算法源于对鸟群觅食和鱼群学习行为的研究,在很多领域得到了广泛应用,本文介绍了粒子群算法的基本原理,并针对粒子群算法在不同应用领域的需求,详细讨论了粒子群算法的各种改进技术,最后,对粒子群算法未来发展进行了展望:粒子群算法;惯性权重;学习因子:TP393:A:10092714(2011)02010405(ParticleSwarmOptimization,PSO)[1,2],1995EberhartKennedy,,,PSO,[3],11.1,itvti=[vti1,vti2,vtid],itxti=(xti1,xti2,xtid),d,pBes,tgBes,tpBest[3],t+1:vt+1id=vtid+c1r1(pBesttid-xtid)+c2r2(gBesttgd-xtid)(1)xt+1id=xtid+vt+1id(2)(1),,c1c2(),r1r2[0,1],,,c1c22:20101113:(T200806)(Q20082203):(1977),,,1041.2PSO[3]:Step1:,pBest,gBes;tStep2:,;Step3:,;Step4:,;Step5:id(1)(2);Step6:,Step2,gBest2,,,2.1PSO,,,,;,,,,,,,;,,[6]1):,,[4];[5]0.90.4,PSO,,,,,;[7],,2):,,::[8],,,,,:[9],:[10]:(t)=(start-end-d1)=e11+d2t/tmax(3),d1,d2,:d1=0.2,d2=0.73):[11],,[12],[13],[14]2.2105PSO,c1c2,c1,;,c2,,;,,[15],[15]RosenbrockRastriginGriewank,c1c2,[16]PSOK,,K,,2.3PSO,PSOPSO,,PSO[17],;[18]PSO,,,,[17,18],[19]PSO,,2.4PSO,PSO,,,[21]1)PSO-DV:[20]PSO(DifferentialOperator),,,2)GA-PSO:,,,,PSO,[21]GAPSO,benchmark17100,GA-PSO,3)SA-PSO:[22]SA-PSO,PSO,,PSOPSO,[22]4,,SA-PSOPSO4)PSACO:[23]PSACO,PSO,PSO,ACO,,65)CPSO:[24],,,,,2.5PSOPSO,PSO,,PSO(DiscretePSO,DPSO)1061)PSO:[25]PSOPSO(BinaryParticleSwarmOptimization,BPSO),DPSO,,1,2)DPSO:,,,,[26]TSP-DPSO3)BPSO:BPSO,,,BPSOPSO[27]PSO(EDPSO),,,;[28],PSO(IPSO),,,[29,30],;[31]PSO,,,,2.6PSO(NichingTechnique),DeJong1975,,[21][32]PSO,PSOPSO,,,,,,,[33]PSO,PSO,,,,,PSOPSO,PSO:[1]KennedyJ,EberhartRC.Particleswarmoptimization[C].ProcIEEEIntCon.fNeuralNetworks.1995(4):1942~1948.[2]EberhartRC,KennedyJ.Anewoptimizerusingparticleswarmtheory[C].inProc6thIntSymp.MicroMachineandHumanScience.1995.39~43.[3],.[M].:,2009[4]YSh,iR.CEberhart.Amodifiedparticleswarmoptimizer[C],inProc.IEEEWorldCongrComputIntel,l1998.69~73.[5]ShiY,CEberhartR.EmpiricalStudyofParticleSwarmOptimization[C].ProceedingofCongressonEvolutionaryComputation.Piscataway,NJ:IEEEServiceCenter.1999.1945~1949.[6],.[M].:,2010.[7],.[J].,2007,33(11),193~195.[8],.[J].,2007,43(23):89~91.[9],.[J].,2007,43(4):47~48.107[10],.[J].,2007,21(4):16~20.[11]ClercM.TheWayoflifeofCheap-PSO,anAdaptivePSO.Technicalreport.[12],.[J].,2006,18(10):2969~2971.[13],.[J].,2007,43(7):68~70.[14]Peng,J,ChenY.BatteryPackStateofChargeEstimatorDesignusingComputationalIntelligenceAproaches[C].InProceedingsoftheAnnualBatteryConferenceonApplicationandAdvances.2000,173~177.[15],.PSO[J].(),2007,4(4):1~4.[16]ClercM.TheWayOfLifeofCheap-PSO,anAdaptivePSO[EB/OL].Technicalreport.[17]KennedyJ,MendesR.PopulationStructureandParticleSwarmPerformance[C].ProcessdingsoftheIEEECongressonEvolutionaryComputation.Piscatawat,2002,NJ:1671~1675.[18]KennedyJ.Stereotyping.ImprovingParticleSwarmPerformancewithClusteranalysis[C].ProceedingsoftheCongressonEvolutionaryComputing.Piscatawat,2000,NJ:1507~1512.[19],.[J].,2007,34(3):205~207,233.[20]DasctalS.ParticleSwarmOptiizationandDifferentialEvolutionAlgorithms:TechnicalAnalysis,ApplicationandHybridizationPrespectives[M].StudiesinComputionalIntelligence.2008.[21][M].:.2010[22],.[J].,2004,(1):47~50.[23]ShelokarPS,PatrickSiarry.Particleswarmandantcolonyalgorithmshybridizedforimprovedcontinuousoptimization[J].AppliedMathematicsandComputation.2007,188:129~142.[24],.[J].,2004,31(8):13~15.[25]KennedyJ,EberhartRC.ADiscreteBinaryVersionoftheParticleswarmAlgorithm[C].The1997ConferenceonSystem,CyberneticsandInformatics,1997,4104~4108.[26]ClertM.DiscreteParticleSwarmOptimization[EB/OL]IllustratedbyTravelingSalesmanProblem.[27],.[J].,2008,36(6):1242~1248[28],.[J].,2008,29(6):1089~1092.[29]SunJ,FengB.ParticleSwarmOptimizationwithParticleHavingQuantuBehavior[C].Proceedingofthe2004IEEECongressonEvolutionaryComputation,2004:325~331.[30]YangS,Wang,M,JiaoL.Aquantum.Particleswarmoptimization[C].Proceedingofthe2004IEEECongressonEvolutionaryComputation,2004,1:320~324.[31],.[J].,2005,26(8):1331~1334.[32]BritsA,PEngelbrecht.AnichingparticleswarmOptimizer[C].InProceedingsoftheConferenceonSimulatedEvolutionandLearning,2002.[33],.[J].,2005,134(34):680~689.ResearchonparticleswarmalgorithmandimprovedtechnologyLIZhi,CHENNiansheng,GUOXiaoshan,KEZongwu(CollegeofComputerScienceandTechnology,HubeiNormalUniversity,Huangshi435002,China)Abstract:Particleswarmoptimization(PSO)isapopulationbasedstochasticoptimizationtechniquedevelopedbyDr.EberhartandDr.Kennedyin1995,inspiredbysocialbehaviorofbirdflockingorfishschooling,itwaswidelyusedinmanyfiled.Thebasicprincipleofparticleswarmalgorithmissimplyintroduced,someimprovedparticleswarmalgorithmsarediscussedandsummarizedinthispaper.Finally,thefutureofparticleswarmoptimizationalgorithmisprospected.Keywords:particleswarmoptimization;inertiaweight;accelerationcoefficients108