欢迎来到天天文库
浏览记录
ID:31977449
大小:2.04 MB
页数:57页
时间:2019-01-29
《基于动态子群策略的改进pso算法与其应用-研究》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、西安理工大学硕士学位论文andshowsitssuperiorityinsolvingpracticalproblems.Onthebasisofintroducingthepowerfulteachingandlearningglobalsearchoperator,westudyanddesignanimprovedparticleswarmoptimizationalgorithmwithanewsubgroupsplittingmethodanddynamicsubgroupstrategy.Wea
2、doptarandomgroupingstrategyinthenewsubgroupsplittingmethod.WenamethenewalgorithmasDynamicSubgroupStrategyParticleSwarmOptimizationalgorithm(DSSPSO).Toexaminetheeffectivenessoftheproposedalgorithm,itistestedon10benchmarkfunctions.Experimentalresultsshowthat
3、DSSPSOisaneffectivemethodtooptimizecomplexfunctions,andperformssignificantlybetterthantheoriginalPSOalgorithmandotherimprovedPSOalgorithmi.e.PSO-wFIPSandUPSO.(1)Inmanypracticaloptimizationproblems,therearesomeconstraintconditions.Swarmintelligencealgorithm
4、shavebeensuccessfullyappliedforconstrainedoptimizationproblems.However,thedifferencebetweeneachconstraintproblemsdifferinthousandsofways,andweusethepenaltyfunctionmethodtosolvetheconstrainedoptimizationproblemisverydifficulttosettheparametersoftheappropria
5、tefactors.Henceweintroduceanewmechanismtoimprovetheparametersconfiguretermeddynamicpenaltymechanismwhichisconducivetosearchforthemobilefeasiblesolutionregiontoacceleratetheprocessofevolution.Basedonthenewproposedmechanismofdynamicpenalty,weapplytheimproved
6、DSSPSOalgorithmtosolvecrudeoilblendingproblemwithmulti-constraintconditions.ThenwegivethemainflowofcrudeoilblendingoptimizationproblemusingdynamicpenaltymechanismandDSSPSO.Finally,weusetwotestdatatotestifythevalidityofthemethod,andthetestresultsshowtheeffe
7、ctivenessofthedynamicpenaltymechanismandDSSPSOalgorithm.Keywords:DSSPSO;DynamicSubgroupStrategy;ConstrainOptimization;DynamicPenaltyii万方数据目录目录1绪论...............................................................................................................
8、.......................................................11.1课题研究背景...............................................................................................................................................
此文档下载收益归作者所有