遗传神经网络的民用航空器SDR预测

南京航空航天大学硕士学位论文基于遗传神经网络的民用航空器SDR预测姓名：解海涛申请学位级别：硕士专业：交通信息工程及控制指导教师：孙建红20090201I100SDRSDRSDRSDRSDRBPSDR73720071-12SDRBPSDRBPBPBPBP737SDRBPSDRBPSDRSDRBPSDRIIAbstractWiththesustainedandrapiddevelopmentofdomesticaviationindustry,thenumberofcargoplanehasbeenincreasedbymorethan100peryearinrecentyears.SecurityisaneternalthemeofCivilAviation;TheresearchofmypapermainlydealswithcivilaircraftSDRprediction,whichisbasedontheanalysistoBPneuralnetworktheory.ItanalyzesmanyfactorsinfluencingSDR,pointsoutthemaladyexistinginusingtraditionalforecastingmethodstodothepredictionontheSDR,andputsforwardanewwaybyusingBPneuralnetworktopredictSDR;throughtheestablishmentofthemodel,itsuccessfullyforecastedtheSDRquantityfromJanuary2007toDecemberontheBoeing737model.AimattheinherentdefectsofSlowconvergenceandeasytofallintolocalminimumpointduringBPneuralnetworkadjustsweights,thepapertriestheapproachofgeneticalgorithmtooptimizetheBPneuralnetwork,doestheempiricalresearchongeneticBPneuralnetwork,andprovesthatthereisagreatimprovementofusingonpredictionaccuracyandrunningvelocity.UsingthismethodtoPredictSDRquantityin2008and2009ontheBoeing737model,ithasgoodapplicabilityandulteriorlyimprovestheaccuracy..Finallycometotheconclusion,geneticBPneuralnetworkinpredictionappliestocivilaircraftSDRhasaverygoodprospectandapplicationvalue.Keywords:SDRpredictNeuralNetworkBPalgorithmGeneticalgorithmGA-BPalgorithm.SDRVI1.1SDR...........................................................................................................32.1.............................................................................................................112.2MP.....................................................................................................................122.3S.......................................................................................................................132.4.................................................................................................................142.5.................................................................................................152.6.............................................................................................................152.7.........................................................................................................172.8.................................................................................................172.9.................................................................................172.10...................................................................................................182.11BP........................................................................................................202.12BP..................................................................................................233.12007...............................................................333.220071-12SDR......................................................................354.1.................................................................................................................425.1BP......................................................................................525.2.........................................................................545.3.........................................................................................................555.4.............................................................................................................................552.1.........................................................................................................143.1MATLAB.........................................................................................293.2BP................................................................................................................323.3B-7371996-20071-12SDR.................................................................343.4.....................................................................................................365.3BPBP...........................................................56VIIwijijMuiiSxjjfiθiHwi(0)()fHHηGABPAverageEMAE()ijwk∆SSEcmMSECPjeMPN()fpECTk11.1200711449.66%20081001500SDRSDR1.2SDRSDRCCAR121707708709135439441443SDRSDR2SDR24SDR72SDRSDRSDR1.11SDR2SDRSDR3SDRSDR43SDRSDR56SDR1.1SDR1.3SDRSDRSDR4SDRSDRSDRSDR1.3.1[3](AR)(MA)(ARMA)(ARIMA)SDRSDRSDRSDRSDRSDR51.3.2[9,37]nn-1()mnijp∧×1.3.380GM(1,1)GM(1,1)GM1,1SDRSDR1.3.41987LapedesFarberSDR6[1,3,30]SDR1p23456BPBP/I/O/I/O[34,35]12BP3SDRSDR1.4SDRBPBP7737SDR1SDR2BP3BP4SDRSDRSDRSDRSDRBPBPBPSDRMATLABMATLABBP7372007112SDRBP737SDRSDR8BPArtificialNeuralNetworks,ANN(NN)[46]2.11943W.S.McCullochW.A.PittsMP6060[3]1.Cajal1943W.S.McCullochW.A.PittsMP1949D.O.HebbHebbHebbHodykinHuxleyNa+K+Hodykin-Huxley92.1958F.RosenblattMarkIPerceptronRosenblattRosenblattB.WidrowAdalineAdalineWidrowRosenblattWidrowK.SteinbuchN.Nilsson19653.1969M.MinskyS.PapertM.Minsky15S.GrossbergT.Kohonen4.1982J.J.HopfieldNPTravellingSalesmanProblemTSPHopfield19