基于遺傳算法的正則化方法研究及應用.pdf

1、湖南大學碩士學位論文基于遺傳算法的正則化方法研究及應用姓名：王貴申請學位級別：碩士專業(yè)：@指導教師：李光耀韓旭20070301基于遺傳算法的正則化方法研究及應用IIAbstractTherearealotofapplicationsofinverseproblemsinscienceengineeringinverseproblemitssolutionmethodshavebeenahotresearchfield.Thediffic

2、ultyinsolvinginverseproblemisitsillposedacteristicthatistheexistence、uniquenessstablilityofthesolution.Fthemostinverseproblemsinengineeringtheexistenceofthesolutionisevidenttheremustbethereasonfakindofresult.Buttheunique

3、nessstablilityofthesolutionaretheproblems.Therehavebeenlotsofcomputationalinversetechniquesfsolvingthiskindofillposedinverseproblems.InallthesemethodsTikhonovregularizationmethodisthemostpopularmethodtosolvetheillposedin

4、verseproblems.(1)TheillposednessistheacteristicofinverseproblemstakeanexampleoftheFredholmintegralequationtoillustratetheinfluenceofillposedness.(2)ThemodifiedTikhonovregularizationmethodispresentedbasedongeicalgithmthes

5、uperiitiesofthetwoalgithmsareinherited.Thenumericalcomputationofthemethodisveryimptanttotheprecisionofthesolutiongeicalgithmisemployedtoimplicittheregularizationalgithmnumericallyasitisakindofadaptiveglobalalgithmbasedon

6、evaluation.Thevalidityofthiscombinedmethodisprovedbytheinverseparameterestimationproblemofacantileverbeam.(3)ThemethodisappliedtoaparameterestimationofadrawbeadmodeltheBauschingereffecttheinclineoftheneutrallayerareestim

7、atedbythemeasureddrawbeadrestrainingfce.Akindofcredibleeffectivestableinversesolutionmethodispresentedfillposedinverseengineeringproblemsinthisresearch.Wecanbetterestimatedthesystemimprovetheefficiencybyimplementingtheme