免疫克隆算法求解动态多目标优化问题

免疫算法

ISSN1000-9825,CODENRUXUEWE mail:jos@iseu.虻∞JournalofSoftwem,Vol18,No.11,November2007,PP.2700-2711

DOI:10136000s182700

o2007byJournalof跏口陀.Alldgh协restr,'edTe]/F&x:+86—10也562563http://www.josorgc“

免疫克隆算法求解动态多目标优化问题’

尚荣华i焦李成,公茂果,马支萍

(西安电子科技大学智能信息处理研究所,陕西西安7100711

AnImmuneClonalAlgorithmforDynamicMulti—ObjecfiveOptimization

SHANGRong-Hua+,JIAOLi-Cheng,GONGMso,Guo,MAWen Ping

(1natlmteofIntelligentInformationProcessing,XidianUniversity,Xi’art710071,China)

+Cerrespondingauthor:Phn:+86-29 88209786,Fax:+86-29-88201023.E mail:thshang@mail.xldianedu.cn’http://www.xidimLeducnShangRIt,JiaoLC,Gong

optimization.JournalMG,MaWP.AnImmuneclonalalgorithmfordynamicmulti-objectivehtm01"Software,2007,18(11):2700—2711.http:llwww.jos.org.cull000-9825/18/2700

ofAbstract:Thediffienlty

oroynminMulti-ObjectiveOptimization(DMo)problem

CIonalAlgorithmliesineithertheobjectivefunetionandconstraiutilnnmnetheassociatedproblemparametersvariationwithtime.Intllispap扎basedo丑theforDMO(ICADMO)is

thePereto-dominanceis

onesclonaltheory,anewDMOalgorithmtermed∞Immnaeproposed.Inthealgorithm,theadopted.Theindividualsinthe

ones,andthehOlrl—dominatedentirecloningisadoptedandtheclonalselectionbasedonantibodyarepopulationmedividedintotwopaas:Dominatedand

Onnon.dominatedoutsselected.Threeoperatorsa咒inUoducedintoICADMO.whichguaranteesthediversi慨theuniformityand

convergencetheeonveTgcnceoftheobteinedsolutions.ICADMomuchiatestedfourDMOtestproblemsandcomparedwiththeanddiversityoftheDirection—BasedMethod(DBM),andobtainedsolutionsisobserved.betterperformanceinboththe

Keywords:artificialimmunesystem;Pareto—optimalfront;dynamicmulti objectiveoptimization;performance

11fietric

摘要:求解动态多目标优化(d”aIllicmulti-objective

法一一动态多目标免疫克隆优化(imm衄eclonaloptimization,简稀D呦)问题的主要困难在于目标函数,约束条件或者相关的问题参数是随时间不断变化的.基于免疫克隆选择学说,提出一种用于解决DMOf,q题的新算algorithmforDMO,简称ICADMo).谊算击改进了现有的克隆策略。采用整体克隆的方式;在选择笨略上,根据Pereto-占优的概念,将抗体群中的个体分为支配个体和非支配个体对非支配个体进行选择,采用3个特色算子,使其很好地保持了所得解的多样性、均匀性和收敛性.通过数值实验,与DBM(direotinn-basedmethod)算法进行比较,结果表明,新算法在收敛性、多样性以及解分布的广度方面都体现了很好的性能.

关键词:人工免疫系统;Pareto.前沿面:动态多目标优化:性能指标

文献标识码:A中国法分类号:TPl8

SupportedbytheNationalNaturalScienceFoundationofChinamade*GrantNos60133010,60372045,60703108(国家自然科学基叠):the

(973))NationalBasicResearchProgramofChinaunderGrantNos2001CB309403,2006CB705700(国家重点基础研究发展计划

Received2006-09-18;Accepted2006-11-21

免疫克隆算法求解动态多目标优化问题相关文档

最新文档

返回顶部