資源描述：
《Genetic Algorithms A survey》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、GeneticAlgorithms:ASurveyM.Srinivas,MotorolaIndiaElectronicsLtd.LalitM.Patnaik,IndianInstituteofSciencenthelastfiveyears,geneticalgorithmshaveemergedaspractical,robustop-timizationandsearchmethods.Diverseareassuchasmusicgeneration,ge-neticsynthesis,VLSItechnology,str
2、ategyplanning,andmachinelearninghaveprofitedfromthesemethods.Thepopularityofgeneticalgorithmsisreflectedinthreebiennialconferences,anewinternationaljournal,andanever-increasingmassofliteraturedevotedtothetheory,practice,andapplicationsofsuchtechniques(seethesidebar“T
3、olearnmore”).Geneticalgorithmsearchmethodsarerootedinthemechanismsofevolutionandnaturalgenetics.Theinterestinheuristicsearchalgorithmswithunderpinningsinnaturalandphysicalprocessesbeganasearlyasthe1970s,whenHolland’firstpro-posedgeneticalgorithms.Thisinterestwasrekin
4、dledbyKirkpatrick,Gelatt,andVecchi’ssimulatedannealingtechniquein1983.2Simulatedannealingisbasedonthermodynamicconsiderations,withannealinginterpretedasanoptimizationpro-Geneticalgorithmscedure.Evolutionary~trategies~.~andgeneticalgorithm^,^.'^ontheotherhand,providea
5、nalternativedrawinspirationfromthenaturalsearchandselectionprocessesleadingtothesur-vivalofthefittestindividuals.Simulatedannealing,geneticalgorithms,andevolu-totraditionaltionarystrategiesaresimilarintheiruseofaprobabilisticsearchmechanismdi-rectedtowarddecreasingco
6、storincreasingpayoff.Thesethreemethodshaveaoptimizationhighprobabilityoflocatingtheglobalsolutionoptimallyinamultimodalsearchtechniquesbyusinglandscape.(Amultimodalcostfunctionhasseverallocallyoptimalsolutionsaswell.)However,eachmethodhasasignificantlydifferentmodeof
7、operation.directedrandomSimulatedannealingprobabilisticallygeneratesasequenceofstatesbasedonasearchestolocatecoolingscheduletoultimatelyconvergetotheglobaloptimum.Evolutionarystrate-giesusemutationsassearchmechanismsandselectiontodirectthesearchtowardtheoptimalsoluti
8、onsinprospectiveregionsinthesearchspace.Geneticalgorithmsgenerateasequenceofpopulationsbyusingaselectionmechanism,andusecrossoveran
