您好,欢迎访问三七文档
当前位置:首页 > 商业/管理/HR > 销售管理 > 基于BP神经网络的供应商评价方法研究
上海交通大学硕士学位论文基于BP神经网络的供应商评价方法研究姓名:成果申请学位级别:硕士专业:项目管理指导教师:苏强;张忠明20090101BPIBPBPBPLevenbergMarquardtBP,BPBPMATLABBPIIRESEARCHONTHEMETHODOFSUPPLIEREVALUATIONBASEDONBPNEURALNETWORKABSTRACTFacingtheeconomicglobalizationandintegration,moreandmorecompaniesareincreasinglythinkingintermsofcompetingaspartofasupplychainagainstothersupplychains,ratherthanasasinglefirmagainstotherindividualcompanies.Selectingagoodsupplierisagoodopportunitytoimprovethesupplychain.Sosupplierevaluationandselectionaretheextremelyimportantdecisiontoeverycompany.Itcouldinfluencethesuccessfulofthecompanyincompetitionorprogramflawlesslaunch.ThisthesisanalyzesthemethodsonthesupplierevaluationbasedonBPNeuralnetwork.Firststepistocreatetheintegratedsystemofsupplierevaluationindexesbycomparingsomesupplierevaluationchecklistsfromautomobilecompanies.SecondoneistocalculatetheweightoneachindexbythemethodofAHP.ThirdstepistocreatethesupplierevaluationmodelbasedontheBPneuralnetworkandtaketheLevenberg-Marquardtoptimizationalgorithmastrainingmethod,whichcouldcausetherapidconvergence.Finally,thetrainedmodelisverifiedbytestsamples.Supplierevaluationisamulti-objectivedecisionproblem.TheresearchBPIIIshowstheBPevaluationmodeldoesbenefittodefinetheweightofnetworkwithdiversityandvariability.TheBPneuralnetworkbasedmodelcouldprovideanoperationalevaluationmethodforsupplierselection.KEYWORDS:SupplierEvaluation,AHP,BPNeuralNetwork,MATLABBPV3-1/BiologicalNerveCellStructure......................................................203-2/NerveCellStructure........................................................................213-3f(a)/f(a)type.......................................................................................................223-4BP/BPNetworkStructure............................................................................233-5/Designofneuralnetworkstructure..............................................294-1/Resultonthenettraining............................................................35BPVI1-1.......................................................................................22-1...............................................................................................62-2...................................................................................................92-3..........................................................................................................................122-4......................................................................................................................142-5..................................................................................................................142-6.........................................................................................................152-7.................................................................................................162-8.................................................................................................162-9.....................................................................................162-10...............................................................................................174-1BP................................................................................................324-2.................................................................................................334-3.............................................................................................344-4.............................................................................344-5.........................................................................................364-6.............................................................................................................36BPBP1BP1.1[1],,;;,;,PMBOK[2]PMBOK[2]1234562[3]1-11-163482783493560484845ENet200711.2,[4]15011028.1%,45.6%,,,3,,31[5]2DEADataEnvelopmentAnalysisDEADEA1978A.CharnesW.W.CooperE.Rhodes3AHP(analytichierarchyprocessAHP)AHPAHP,4BP1.3BP,,,,BPBPLevenbergMarquardtBP,BPBPBPLevenbergMarquardtBP(TrytoTry)BPBPMATLABBPBP51.43DEAAHPBPBP6,,,,,2.12.1.12-12-15005000-90%300010000-70%550010000-45%10001800-44%1400022000-36%WallStreetJournal(August16,2004):B1andProductionandInventoryManagementJournal(ThirdQuarter2004:18-22)1234,,617234562.1.2GWDickson17023Dickson14ISO/TS169495SDickson38,2-2,5152.1.3,,,,(),(),::)/()(minmaxminXXXXYii−−=(2-1):)/()(minmaxmaxXXXXYii−−=(2-2):iX,maxX,minX,iY[0,1]2.1.4(SPI)2-2,,(SPI),,:∑==niiiYWSPI1)((2-3):iW,iY,(2-1)(2-2)92-2102.2AHPAHP,,,2.2.1AHPTheanalytichierarchyprocessAHPThomasL.Staaty[6]701971AHP1977AHPAHPAHPAHPAHP1982[9]H.Gholamnezhad1982AHP19879AHP1988AHPAHPAHP[7]2.2.211AHP(1):AHP123(2):1nnww,,1Λnn},,{1nxxXΛ=ZSaaty[13]ixjxijaixjxZnnijaA×=)(AXZ−ixjxZijajxixZijjiaa1=nnijaA×=)(i0ijaiiijjiaa1=nji,,2,1,Λ=(1=iiani,,1Λ=)ijaSaaty19[11]19122-3135792468ijijajiijjiaa1=Saaty192)1(−nn1−n2)1(−nn(3):AmaxλWAikjkijaaa=nkji,,2,1,,Λ=∀2-42-1AAAmaxλAmaxλA13iAiiATAiiiA1)(rank=AAivAn=maxλnAAvAmaxλTn),,(1Λ=jiijwwa=nji,,2,1,Λ=∀=n
本文标题:基于BP神经网络的供应商评价方法研究
链接地址:https://www.777doc.com/doc-554353 .html