本文關(guān)鍵詞：RBF神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)動(dòng)態(tài)優(yōu)化設(shè)計(jì)，由筆耕文化傳播整理發(fā)布。

圖8Fig.8

D-RBF神經(jīng)網(wǎng)絡(luò)訓(xùn)練過(guò)程Thetrainingprocessof

D-RBF

圖10Fig.10

基于D-RBF的COD建模結(jié)果

ThemodellingofCODbasedon

D-RBF

圖9

Fig.9

訓(xùn)練過(guò)程中神經(jīng)元數(shù)

Fig.11

圖11基于D-RBF的COD建模誤差

Theneuronsleftinthetrainingprocess

ThemodellingerrorsofCODbasedonD-RBF

推薦值.訓(xùn)練過(guò)程中誤差變化如圖8所示,訓(xùn)練過(guò)程中隱含層神經(jīng)元的變化如圖9所示,對(duì)COD的建模結(jié)果如圖10和圖11所示.

仿真結(jié)果表明:原水中的有機(jī)污染物(COD約300～500mg/L)得到有效去除(出水負(fù)荷COD在不同的時(shí)刻都能保持在30mg/L左右),圖10和圖11顯示實(shí)測(cè)COD值與本模型的輸出值基本吻合,相對(duì)誤差小于0.02,證明該模型是有效的.1)不依賴于RBF網(wǎng)絡(luò)的初始結(jié)構(gòu),能夠根據(jù)實(shí)際對(duì)象調(diào)整RBF神經(jīng)網(wǎng)絡(luò),獲得適合對(duì)象的RBF神經(jīng)網(wǎng)絡(luò);

2)通過(guò)神經(jīng)網(wǎng)絡(luò)輸出敏感度增加和刪除RBF神經(jīng)網(wǎng)絡(luò)中的神經(jīng)元,最終獲得的RBF神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu)簡(jiǎn)潔,逼近能力強(qiáng);

3)本文提出的D-RBF神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu)穩(wěn)定,為復(fù)雜系統(tǒng)建模提供了技術(shù)支持.

References

1YeJian,GeLin-Dong,WuYue-Xian.Anapplicationofim-provedRBFneuralnetworkinmodulationrecognition.ActaAutomaticaSinica,2007,33(6):652?654

(葉健,葛臨東,吳月嫻.一種優(yōu)化的RBF神經(jīng)網(wǎng)絡(luò)在調(diào)制識(shí)別中的應(yīng)用.自動(dòng)化學(xué)報(bào),2007,33(6):652?654)2ChenS,WangXX,BrownDJ.Sparseincrementalre-gressionmodelingusingcorrelationcriterionwithboost-ingsearch.IEEESignalProcessingLetters,2005,12(3):198?2013ChenS,WolfgangA,HarrisCJ,HanzoL.SymmetricRBF

3結(jié)論

針對(duì)RBF神經(jīng)網(wǎng)絡(luò)設(shè)計(jì)結(jié)構(gòu)問(wèn)題,提出了一種基于輸出敏感度法的動(dòng)態(tài)RBF神經(jīng)網(wǎng)絡(luò)(D-RBF),D-RBF在保證神經(jīng)網(wǎng)絡(luò)收斂性能的前提下實(shí)現(xiàn)結(jié)構(gòu)在線調(diào)整,提高神經(jīng)網(wǎng)絡(luò)的自適應(yīng)能力;通過(guò)逼近非線性函數(shù)和對(duì)非線性系統(tǒng)關(guān)鍵參數(shù)進(jìn)行建模,以及與其他動(dòng)態(tài)RBF神經(jīng)網(wǎng)絡(luò)進(jìn)行比較,得到以下結(jié)論

:

classi?erfornonlineardetectioninmultiple-antenna-aidedsystems.IEEETransactionsonNeuralNetworks,2008,19(5):737?745

4ChaiTian-You,WangXiao-Bo.ApplicationofRBFneuralnetworksincontrolsystemoftheslabacceleratingcoolingprocess.ActaAutomaticaSinica,2000,26(2):219?225(柴天佑,王笑波.RBF神經(jīng)網(wǎng)絡(luò)在加速冷卻控制系統(tǒng)中的應(yīng)用.自動(dòng)化學(xué)報(bào),2000,26(2):219?225)5GonzalezJ,RojasI,OrtegaJ,PomaresH,FernandezFJ,DiazAF.Multi-objectiveevolutionaryoptimizationofthesize,shape,andpositionparametersofradialbasisfunctionnetworksforfunctionapproximation.IEEETransactionsonNeuralNetworks,2003,14(6):1478?14956LeungFHF,LamHK,LingSH,TamPKS.Tuningofthestructureandparametersofaneuralnetworkusinganimprovedgeneticalgorithm.IEEETransactionsonNeuralNetworks,2003,14(1):79?887BorsAG,PitasI.Medianradialbasisfunctionneuralnet-work.IEEETransactionsonNeuralNetworks,1996,7(6):1351?13648YinH,AllinsonNM.Self-organizingmixturenetworksforprobabilitydensityestimation.IEEETransactionsonNeu-ralNetworks,2001,12(2):405?4119EspositoA,MarinaroM,OricchioD,ScarpettaS.Ap-proximationofcontinuousanddiscontinuousmappingsbyagrowingneuralRBF-basedalgorithm.NeuralNetworks,2000,13(6):651?66510FuXJ,WangLP.Datadimensionalityreductionwithap-plicationtosimplifyingRBFnetworkstructureandimprov-ingclassi?cationperformance.IEEETransactionsonSys-tems,Man,andCybernetics,PartB:Cybernetics,2003,33(3):399?40911LuYW,SundararajanN,SaratchandranP.Asequential

learningschemeforfunctionapproximationusingminimalradialbasisfunction(RBF)neuralnetworks.NeuralCom-putation,1997,9(2):461?47812PanchapakesanC,PalaniswamiM,RalphD,ManzieC.Ef-fectsofmovingthecentersinanRBFnetwork.IEEETrans-actionsonNeuralNetworks,2002,13(6):1299?130713HuangGB,SaratchandranP,SundararajanN.Ageneral-izedgrowingandpruningRBF(GGAP-RBF)neuralnet-workforfunctionapproximation.IEEETransactionsonNeuralNetworks,2005,16(1):57?6714FengHM.Self-generationRBFNsusingevolutionalPSO

learning.Neurocomputing,2006,70(1-3):241?25115HuangDS,DuJX.Aconstructivehybridstructureop-timizationmethodologyforradialbasisprobabilisticneu-ralnetworks.IEEETransactionsonNeuralNetworks,2008,19(12):2099?2115

16ChenS,HongX,LukBL,HarrisCJ.Constructionoftun-ableradialbasisfunctionnetworksusingorthogonalforwardselection.IEEETransactionsonSystems,Man,andCyber-netics,PartB:Cybernetics,2009,39(2):457?46617SaltelliA,RattoM,TarantolaS,CampolongoF.Sensitivity

analysispractices:strategiesformodel-basedinference.Re-liabilityEngineeringandSystemsSafety,2006,91(10-11):1109?1125

18RattoM,TarantolaS,SaltelliA.Sensitivityanalysis

inmodelcalibration:GSA-GLUEapproach.ComputerPhysicsCommunications,2001,136(3):212?22419CariboniJ,GatelliD,LiskaR,SaltelliA.Theroleofsensi-tivityanalysisinecologicalmodeling.EcologicalModelling,2007,203(1-2):167?18220CampolongoF,CariboniJ,SaltelliA.Ane?ectivescreening

designforsensitivityanalysisoflargemodels.Environmen-talModellingandSoftware,2007,22(10):1509?151821BuzziC,GrippoL,SciandroneM.Convergentdecomposi-tiontechniquesfortrainingRBFneuralnetworks.NeuralComputation,2001,13(8):1891?192022LiuYY,StarzykJA,ZhuZ.Optimizedapproximational-gorithminneuralnetworkswithoutover?tting.IEEETrans-actionsonNeuralNetworks,2008,19(6):983?99523HenzeM,GradyCPL,GujerW,MaraisGVR,Matsuo

T.Ageneralmodelforsingle-sludgewastewatertreatmentsystems.WaterResearch,1987,21(5):505?515

喬俊飛北京工業(yè)大學(xué)教授.主要研究方向?yàn)橹悄芸刂?神經(jīng)網(wǎng)絡(luò)分析與設(shè)計(jì).E-mail:junfeq@bjut.edu.cn

(QIAOJun-FeiProfessoratBeijingUniversityofTechnology.Hisresearchinterestcoversintelligentcontrol,andanalysisanddesignofneuralnetworks.)

韓紅桂北京工業(yè)大學(xué)博士研究生.主要研究方向?yàn)閺?fù)雜過(guò)程建模與控制,神經(jīng)網(wǎng)絡(luò)分析與設(shè)計(jì).本文通信作者.E-mail:isibox@sina.com

(HANHong-GuiPh.D.candidateatBeijingUniversityofTechnology.Hisresearchinterestcoversmodelingandcontrolincomplexprocess,and

analysisanddesignofneuralnetworks.Correspondingau-thorofthispaper.)

博泰典藏網(wǎng)btdcw.com包含總結(jié)匯報(bào)、表格模板、計(jì)劃方案、初中教育、外語(yǔ)學(xué)習(xí)、農(nóng)林牧漁、教學(xué)研究、人文社科、黨團(tuán)工作以及RBF神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)動(dòng)態(tài)優(yōu)化設(shè)計(jì)_喬俊飛_圖文等內(nèi)容。

本文共2頁(yè)12

  本文關(guān)鍵詞：RBF神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)動(dòng)態(tài)優(yōu)化設(shè)計(jì)，由筆耕文化傳播整理發(fā)布。