abstract differential evolution and non-separability using selective pressure to focus sear

abstract differential evolution and non-separability using selective pressure to focus sear

ID:33941945

大小:1.66 MB

页数:8页

时间:2019-02-28

abstract differential evolution and non-separability using selective pressure to focus sear_第1页
abstract differential evolution and non-separability using selective pressure to focus sear_第2页
abstract differential evolution and non-separability using selective pressure to focus sear_第3页
abstract differential evolution and non-separability using selective pressure to focus sear_第4页
abstract differential evolution and non-separability using selective pressure to focus sear_第5页
资源描述:

《abstract differential evolution and non-separability using selective pressure to focus sear》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、DifferentialEvolutionandNon-separability:UsingselectivepressuretofocussearchAndrewM.SuttonMonteLunacekL.DarrellWhitleyDepartmentofComputerScienceColoradoStateUniversityFortCollins,CO80523{sutton,lunacek,whitley}@cs.colostate.eduABSTRACTtor.Thisdescribesapotentialhyper-rectangle,wit

2、htheindividualanditsmutantvectorinopposingcorners,onRecentresultsshowthattheDifferentialEvolutionalgo-whichthenewindividualwilllie.Ofcourse,thesidesofrithmhassignificantdifficultyonfunctionsthatarenotlin-thishyper-rectangleareorientedwiththecoordinateaxes,earlyseparable.Onsuchfunctions

3、,thealgorithmmustrelymakingsearchstepsgeneratedbycrossoverorthogonaltoprimarilyonitsdifferentialmutationprocedurewhich,un-thefunction’scoordinatedirections.Thisoperationiscon-likeitsrecombinationstrategy,isrotationallyinvariant.WetrolledbythecrossoverratevalueCRwhichspecifiestheconje

4、cturethatthismutationstrategylackssufficientselec-expectednumberofparametersinheritedfromthemutanttivepressurewhenappointingparentanddonorvectorstovector.Essentially,CRdeterminestheexpecteddimensionhavesatisfactoryexploitativepoweronnon-separablefunc-ofthesubspaceinwhichtheoffspringwi

5、lllie.tions.Wefindthatimposingpressureintheformofrank-Recentresults[10,3]suggestthatDEstruggleswithfunc-baseddifferentialmutationresultsinasignificantimprove-tionsthatarenotlinearlyseparable.Weconjecturethatmentofexploitationonrotatedbenchmarks.DE’sbiastowardseparabilityarisesfromineffi

6、cientexploita-tionduringthedifferentialmutationstep.InthispaperweCategoriesandSubjectDescriptorswillexplorethefollowingtwohypotheses.I.2.8[ArtificialIntelligence]:ProblemSolving,Control1.WhenCRislow,DEcanexploittheseparabilityofMethods,andSearch;G.1.6[NumericalAnalysis]:Opti-afunctio

7、nbyrelyingoncrossovertoproduceaxis-mization—GlobalOptimizationorthogonalsteps.Onnon-separablefunctions,itmustdependmoredirectlyondifferentialmutation.GeneralTerms2.DElacksadequateselectivepressureinthedifferen-Performancetialmutationsteptomakeefficientprogressonnon-separablefunctions.K

8、eywordsOnfunctionswithsign

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。