In spite of the recent quick growth of the Evolutionary Multi-objective Optimization (EMO) research field, there has been few trials to adapt the general variation operators to the particular context of the quest for the Pareto-optimal set. The only except
5
2
y
a
M
9
2
]
I
A.
s
c
[
1
v
8
5
5
/0
sc
:viXraDominanceBasedCrossoverOperatorforEvolutionaryMulti-objectiveAlgorithmsDRAFTVERSIONOlgaRudenko,MarcSchoenauerTAOTeam,INRIAFutursLRI,bat.490,Universit´eParis-Sud91405OrsayCedex,FranceOlga.Roudenko@lri.fr,Marc.Schoenauer@inria.frAbstract.InspiteoftherecentquickgrowthoftheEvolutionaryMulti-objectiveOptimization(EMO)research eld,therehasbeenfewtrialstoadaptthegeneralvariationoperatorstotheparticularcontextofthequestforthePareto-optimalset.Theonlyexceptionsaresomematingrestrictionsthattakeinaccountthedistancebetweenthepotentialmates–butcon-tradictoryconclusionshavebeenreported.Thispaperintroducesapar-ticularmatingrestrictionforEvolutionaryMulti-objectiveAlgorithms,basedontheParetodominancerelation:thepartnerofanon-dominatedindividualwillbepreferablychosenamongtheindividualsofthepop-ulationthatitdominates.CoupledwiththeBLXcrossoveroperator,twodi erentwaysofgeneratingo springareproposed.Thisrecombina-tionschemeisvalidatedwithinthewell-knownNSGA-IIframeworkonthreebi-objectivebenchmarkproblemsandonereal-worldbi-objectiveconstrainedoptimizationproblem.AnaccelerationoftheprogressofthepopulationtowardtheParetosetisobservedonallproblems.IntroductionTheideaofrestrictedmatingisnotnewinEvolutionaryComputation:Gold-berg[7]alreadysuggestedtoforbid,oratleastrestrict,thecrossoverbetweentoodi erentindividuals(i.e.thataretoofarapartforsomedistanceonthe
genotypicspace)–whichmakessenseforsingle-objectiveproblemsassoonasthepopulationhasstartedtoaccumulateondi erent tnesspeaks,asrecom-biningindividualsfromdi erentpeakswouldoftenleadtolethalindividuals.ThisideahasbeentransposedintheframeworkofEvolutionaryMulti-objectiveAlgorithms(EMAs)byHajelaandLinandbyFonsecaandFleming[6].Nev-ertheless,ZitzlerandThieledidnotobserveanyimprovementwhenmatingsimilarindividuals.Ontheotherhand,Hornetal.presentanargumentsupportingmatingofdissimilarindividuals:inthemulti-objectiveframework,becausethepopulationdiversityisenforced,theinformationprovidedbyverydi erentsolutionscanbecombinedinsuchwaythatanewtypeof(good)com-promisescanhopefullybeobtained.Nevertheless,Scha erreportedtheabsence
evolutionary algorithm for solving multi-objective ...Operator Design 77 Structures of Various MOEAs 88...Dominance Relations 5.4.4 Primary Quality ...
Dominance based crossove... 暂无评价 11页 免费 Hybridized Crossover-Bas..... A new representation and crossover operator for search-based optimization of...
crossover and mutation, the next generation is ...Multi-objective Optimization Evolutionary Algorithm,”... Dominance based crosso... 暂无评价 11页 免费 ...
Dominance based crossove... 暂无评价 11页 免费... One of the issues in evolutionary algorithms (...If we wish to reward a crossover operator for ...
An evolutionary algorith... 13页 1财富值 An ...De?nition 1 (Pareto dominance) An objective ...g. uniform order based crossover, order based ...
cooperative coevolutionary and ε-dominance based ...Uniform distribution mutation operator is adopted to...Evolutionary multi-objective optimization: A ...
Controlling dominance ... 16页 免费 Non-uniform...C. Crossover Operator For a tradeoff accelerating...multi-objective evolutionary algorithms,” in Proc....
(2002) proposed a dominance-based selection scheme...Secondly, the crossover operator is implemented in...L. (1999). Evolutionary techniques for constrained...
Crossover in Evolutionary Multiobjective Optimization ...The r-Dominance: A New Dominance Relation for ...Using a Neighborhood-Based Mutation Operator Das, ...
selection and the concept of Pareto dominance. ...erent multi-objective evolutionary algorithms by ...This process represents the crossover operator in ...
热门文档