基于分形的全局优化算法与其应用

《基于分形的全局优化算法与其应用》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、AGlobalOptimizationFractal．．basedAlgorithmandItsApplicationABSTRACTTheglobaloptimizationproblemisataskwhichhasbeenresearclledandstudiedformanyyears．Itiscommonlyfoundinallareasofreallife．suchasfinance，aerospace，civilengineering．Thedemandsforoptimizingismoreandmorehigh，andmanytraditionaloptimizatio

2、nmethodsaredifficulttoadapttheseproblemswhicharediscontinuous，no—linear，orwithhighdimension，soseekingnewoptimizationmethodhasbeenanimportantresearchdirectionofmanyfields．Atpresentthemethodsofsolvingoptimizationproblemsaredividedintotwokinds．deterministicalgorithmandrandomalgorithm，whichhavediffer

3、entscopeofapplicationaccordingtothecharacteristicsofproblems．Generally，randomalgorithmissomekindofsimulationofnaturalphenomenon．Its‘demandforthepropertiesofproblems’objectivefunctionisverylow，evennodemandsometimes，SOitcansolvethosehardproblems．Heuristicalgorithmisonekindofrandomalgorithm，anditsgo

4、odapplicationspeeduppeople’Sstepstoexplorevariousoptimizationmethod．Inrecentyears，somescholarsusedffactalintheareaofoptimization．andattemptedtodealwithcomplexproblemsusingfractal．Fractalalgorithmisakindofdeterministicalgorithm．itsearchforoptimalsolutionbydividingthefeasibleregioniteratively．Itiso

5、nlyappropriateforsimpleproblemswithlowdimension．Forhighdiinensionalandcomplexproblemsintoday’Ssocietv．itisalmostpowerless．Inordertobreakthissituation．thispaperproposesanimprovedfractaloptimizationalgorithm，thisalgorithmsimulatesthenature”fractal2rowth”phenomenon，changestheoriginalalgorithmsegment

6、ationstrategycompletely．andbringsInagrowthhormonefactortoaccelerateconvergencespeed．Afterthreeclassicfunctionoptimizationexperiment．wecanfindthatfractaloptimizationalgorithmisbetterthantheoriginalmethod，andismoresuitableforproblenlswithhighdimensional．Finally．wetrytoresolvecombinationaloptimizati

7、onDroblenlusingthisalgorithm．Theactualproblemismorecomplexthansimplefunctionformula，anditsjsolutionspaceisdiscrete．thefeasibleregionneedroOFesDecialprocessing．Whenapplythisalgorithminthetravelingsalesmanproblem,results