遺傳編程的改進及其在符號回歸中的應用.pdf

1、長沙理工大學碩士學位論文遺傳編程的改進及其在符號回歸中的應用姓名：潘小海申請學位級別：碩士專業(yè)：計算機應用技術指導教師：徐蔚鴻20100301ABSTRACTRegressionanalysisisoneofthemostimportanttasksinmeasurementdataprocessingThemeaningofsymbolicregressionisthatwearesearchingforafunctionthatcl

2、oselymatchesanunknownexpressionbasedonafinitesetofsampledata，inordertoanalyzeandforecastdataCommonlyusedmethodisprovidingempiricalexpressionmodelandestimatingtheregressionparametersThefittingresultsarebadwhenthefunctionm

3、odelsarehardtobeprovidedGeneticprogrammingisanewtechnologyforsolvingsymbolicregressionproblemItcansolvecomplexregressionproblembecauseitcanobtainthematchedfunctionexpressionbyonlygiventhedatapointsandacceptableerrorThema

4、inresearchcontributionofthethesiscanbesummarizesasfollows：(1)Anovelgeneticprogramming，namedsymbolgeneticprogramming(SGP)basedonanewencodingmethodhasbeenproposedThisnewencodingmethodabsorbsthemeritsofmanyotherlineargeneti

5、cprogrammingsItcodeswithasimple，unrestrainedstring，basedonitscharacters，multipleexpressionscouldbecontainedinoneindividualwithouttheincreaseofcomputationtaskThismethodisprovedtobeeffectiveandstablefromthecomplexityanalys

6、isandexperiment(2)ByintroducingCellularAutomata，anewgeneticprogrammingmethodnamedCellularSymbolGeneticProgramming(CSGP)hasbeenproposedIthashigherSuccessrateandneedlesscomputationtimeThismethodisprovedtobeeffectiveandstab