资源描述：
《penalty methods in genetic algorithm for solving numerical constrained optimization problems》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、PENALTYMETHODSINGENETICALGORITHMFORSOLVINGNUMERICALCONSTRAINEDOPTIMIZATIONPROBLEMSATHESISSUBMITTEDTOTHEGRADUATESCHOOLOFAPPLIEDSCIENCESOFNEAREASTUNIVERSITYbyMAHMOUDK.M.ABURUBInPartialFulfillmentoftheRequirementsfortheDegreeofMasterofScienceinComputerEngineerıngNICOSIA201255Iherebydeclarethatallin
2、formationinthisdocumenthasbeenobtainedandpresentedinaccordancewithacademicrulesandethicalconduct.Ialsodeclarethat,asrequiredbytheserulesandconduct,Ihavefullycitedandreferencedallmaterialandresultsthatarenotoriginaltothiswork.Name,Lastname:MAHMOUDABURUBSignature:Date:55ABSTRACTOptimizationisacomp
3、uterbasedormathematicalbasedprocessusedtofindthebestsolutionincomplicatedhyperspace.Optimizationisanimportantthemethatcanbeusedtoenhanceagivenresultor,toproveit.However,itisproportionaltotheformulationoftheprobleminhand.Optimizationisreallysimpleforsomesortofproblems,butitwillbemorecomplicatedin
4、constrainedhyperspace,whereequalityandinequalityconstraintsexist.EvolutionaryAlgorithmsareoneofthemostpowerfuloptimizationmethodsusedformanytypesofproblems.Geneticalgorithms,otherstrategiesinuse,arealsopowerfuloptimizationtools,astheyarenotinterferedwithbythecomplexityofhyperspace.Ontheotherhand
5、,theyonlyinterferewithtraitsneedtobeoptimizedbymimickingnaturalselectionandenvironmentaladaptationlikegeneticdevelopmentsprocessofanyspecies.Combininggeneticalgorithmswithoptimizationinconstraintshyperspaceisonlybyapplyingpenaltyfunctions.Iftwotypesofconstraintsareon,equality,andinequalityconstr
6、aints;convertingequalityconstraintstoinequalityformatcanbedonebysubtractingaconstantfromconstraintvalue,oftenarationalnumber.Thesatisfactionofconstraintsisthebasicconditionforsolutiontoberecognizedasvalidone.Nevertheless,notallformulatedproblemswillbesolvedbyusinganoptimizationmethod,astheycould
7、sufferfromamisunderstandingoftheproblemortheconstraintsviolations.Thisstudyfocusesonapplyinggeneticalgorithmstoconstraintsproblemsbyapplyingpenalty.Threetypesofalgorithmsareused,dynamicpenalty,staticpenaltyandstochasticranki
