人工神经网络在ERP系统中的应用

:2004-09-24:(1979-),,,,:;(1959-),,,,:.:1001-9081(2005)04-0748-02ERP1,2(1.,123000;2.,123000)([e-mail]):ERP,,,ERP,,ERPBP,,,,::ERP;;BP;:TP182:AApplicationinERPsystemforartificialneuralnetworkZHAOYun2feng1,LIUWan2jun2(1.DepartmentofElectronicsandInformationEngineerning,LiaoningTechnicalUniversity,FuxinLiaoning123000,China;2.CollegeofSoftwareEngineerning,LiaoningTechnicalUniversity,FuxinLiaoning123000,China)Abstract:AnexpertsystemmodulewhichisapredictionanalysismodulebasedontheartificialneuralnetworkwasaddedintotraditionalERPsystemforthecomplexnonlinearprediction,sothattheintelligentandautomaticlevelofERPsystemwashigher.ThemodulecarriedoutthenonlinearpredictionbyusingtheBPneuralnetwork.TheBPneuralnetworkgottheinherentrelationbetweentheinputsandoutputsandthesolutionoftheinvolvedproblembyadjustingtheweightoftheneuralnetworkonlineandtraining.Acarmarketofthecarmanufacturingcompanywaspredicted.Theresultshowsthatthemethodpossessestheexcellentperformanceandpracticability,whichcanbeusedasaeffectivemethodforERPsystem.Keywords:ERPsystem;artificialneuralnetwork;modelofbackpropagationnetwork;marketERP(EnterpriseResourcePlanning)[1],,[2],ERP,,,,[3],,,[4],[5]:1),,,;2),,,ERP,,,,,[7],ERP1,,,,,ERP,11ERPERP,,ERP,ERP,,,,25420054ComputerApplicationsVol.25No.4Apr.2005©1995-2005TsinghuaTongfangOpticalDiscCo.,Ltd.Allrightsreserved.2BP[8]BP,BPBP,,,,,BP(),,,,:(1):Wij,Vjk,AjBkWijij;Vjkjk;Ajj;Bkk(2):X={x1,x2,,xi},i(3):IHj=iWijxi+Aj(1)OHj=(1+e-IHj)-1(2)IHj,OHjj(4):IOk=jVjkOHj+Bk(3)OOk=(1+e-IOk)-1(4)IOk,OOkk(5):dk=(OOk-Tk)OOk(1-OOk)(5)j=kdkVjkOHj(1-OHj)(6)dkk;jj;Tk(6)Wij,Vjk,AjBk:Vjk=Vjk+dkOHj(7)Bk=Bk+dk(8)Wij=Wij+jxi(9)Aj=Aj+j(10),,:01,01(7),,(3);(8)(8):Etd=j0.5(OOk-Tk)2=LELL=1,2,,N(11)ELL;L=1,2,,N;N;EtdEtdE(E),ELEe(L=1,2,,N),EeL,(9);(3)(6),(8)(9)Wij,Vjk,AjBk,3,ERP6,3.1,,,5,5,5,1,22BP3.2,,[0,1],X,XX={,,,,}={x1,x2,x3,x4,x5},0x1,x2,x3,x4,x51,8x,,Wij,VjkAj,Bk,8,8118()()(%)3.71593.80882.51.50531.43304.81.28661.27051.33.38633.36530.64.13904.00883.13.02163.10802.93.97253.60919.15.78485.74120.84,,,,9474:ERP©1995-2005TsinghuaTongfangOpticalDiscCo.,Ltd.Allrightsreserved.:1001-9081(2005)04-0750-04,,(,510632)([e-mail]):,,Matlab,,BP,:;;:TP391.41:AAdaptivefuzzyclassifierbasedontable2lookinglearningalgorithmHUANGZhan,JIANGYu2ying,ZHANGLei(DepartmentofComputerScience,JinanUniversity,GuangzhouGuangdong510632,China)Abstract:Animprovedadaptivefuzzysystembasedontable2lookinglearningalgorithm,adaptivefuzzyclassifierwasdeveloped.Forthepatternrecognitionproblems,simulationstudiesweremadebyapplyingtheadaptivefuzzyclassifierandtheBP32layerfeed2forwardneuralnetworkclassifiertothehandwrittendigitrecognitionproblems.AscomparedtotheBPneuralnetwork,theadaptivefuzzyclassifierhasbetterperformanceinrecognitionability,incorporationgthelinguisticinformationandcomptationalsimplicity.Alltheseshowthesuperiorityandpotentialofadaptivefuzzytechniquesinsolvingpatternrecognitionproblems.Keywords:handwrittendigitrecognition;adaptivefuzzyclassifier;artificialneuralnetwork0[2],,,,,,[2],,,,,Matlab,,[1,3,4],[1],BPBP,11.11,,,:2004-09-06;:2004-12-24:(974225);(476):(1965-),(),,,,:;(1965-),,,,,:;(1972-),,,,:.ERP,:[1].ERP[M].:,2002.[2]HALLR,ANDRIANIP.Analysingintangibleresourceandmanaginginasupplychaincontext[J].EourpeanManagementJournal,1998,16(6):685-697.[3]VENUGOPALV,BAETSW.NeuralNetworkandTheirApplicationsinMarketingManagement[J].JouranlofSystemManagement,1994,45(9):16-21.[4]YOONY,SWALESG,MARGAVIOTM.AComparisonofDiscrimi2nantAnalysisVersusArtificialNeuralNetworks[J].JournaloftheOperationalResearchSociety,1993,44(1):51-60.[5],.[M].:,1993.73-88.[6],.[M].:,2001.324-325.[7].[M].:,2001.110-115.[8].[M].:,1995.57-62.25420054ComputerApplicationsVol.25No.4Apr.2005©1995-2005TsinghuaTongfangOpticalDiscCo.,Ltd.Allrightsreserved.