资源描述：
《免疫克隆算法求解动态多目标优化问题》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。
1、ISSN1000-9825,CODENRUXUEWE-mail:jos@iscas.ac.cnJournalofSoftware,Vol.18,No.11,November2007,pp.2700−2711http://www.jos.org.cnDOI:10.1360/jos182700Tel/Fax:+86-10-62562563©2007byJournalofSoftware.Allrightsreserved.∗免疫克隆算法求解动态多目标优化问题+尚荣华,焦李成,公茂果,马文萍(西安电子科技大学智能信息处理研究所,陕西西安7100
2、71)AnImmuneClonalAlgorithmforDynamicMulti-ObjectiveOptimization+SHANGRong-Hua,JIAOLi-Cheng,GONGMao-Guo,MAWen-Ping(InstituteofIntelligentInformationProcessing,XidianUniversity,Xi’an710071,China)+Correspondingauthor:Phn:+86-29-88209786,Fax:+86-29-88201023,E-mail:rhshang@mai
3、l.xidian.edu.cn,http://www.xidian.edu.cnShangRH,JiaoLC,GongMG,MaWP.Animmuneclonalalgorithmfordynamicmulti-objectiveoptimization.JournalofSoftware,2007,18(11):2700−2711.http://www.jos.org.cn/1000-9825/18/2700.htmAbstract:ThedifficultyofDynamicMulti-ObjectiveOptimization(DM
4、O)problemliesineithertheobjectivefunctionandconstraintortheassociatedproblemparametersvariationwithtime.Inthispaper,basedontheimmuneclonaltheory,anewDMOalgorithmtermedasImmuneClonalAlgorithmforDMO(ICADMO)isproposed.Inthealgorithm,theentirecloningisadoptedandtheclonalselec
5、tionbasedonthePareto-dominanceisadopted.Theindividualsintheantibodypopulationaredividedintotwoparts:Dominatedonesandnon-dominatedones,andthenon-dominatedonesareselected.ThreeoperatorsareintroducedintoICADMO,whichguaranteesthediversity,theuniformityandtheconvergenceoftheob
6、tainedsolutions.ICADMOistestedonfourDMOtestproblemsandcomparedwiththeDirection-BasedMethod(DBM),andmuchbetterperformanceinboththeconvergenceanddiversityoftheobtainedsolutionsisobserved.Keywords:artificialimmunesystem;Pareto-optimalfront;dynamicmulti-objectiveoptimization;
7、performancemetric摘要:求解动态多目标优化(dynamicmulti-objectiveoptimization,简称DMO)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DMO问题的新算法——动态多目标免疫克隆优化(immuneclonalalgorithmforDMO,简称ICADMO).该算法改进了现有的克隆策略,采用整体克隆的方式;在选择策略上,根据Pareto-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
8、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(direction-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
