隨機時滯神經(jīng)網(wǎng)絡模型的數(shù)值解.pdf

1、寧夏大學碩士學位論文隨機時滯神經(jīng)網(wǎng)絡模型的數(shù)值解姓名：岳紅格申請學位級別：碩士專業(yè)：運籌學與控制論指導教師：張啟敏201105寧夏大學碩士學位論文英文摘要AbstractInrecentyearsthestudyofstochasticdelayneuralwks(SDNNs)havebeenbeingattractedwideattention.IngeneralthemajityofSDNNshavenoexactsolutions

2、sodothestochasticdelayneuralwkswithjumps.TherefenumericalapproximationsuchastheEulerschemeisavitaltoolinexplingtheirsolutionsproperties.Themainpurposeofthispaperistoinvestigatetheconvergenceofstochasticdelayneuralwkswith

3、PoissonjumpsMarkovianswitchingthefollowingarethemaincontents:1.AccdingtotheEulermethodweusemartingaleinequalityIt?ofmulathenumericalapproximationfaclassofstochasticdelayLotkaVolterramodelisdiscussed.Inthelastsectionanume

4、ricalexampleisgiven.2.StochasticdelayneuralwkswithPoissonjumponeoftheimptantclassesofhybridsystemsisstudied.AccdingtotheEulermethodweusemartingaleinequalityIt?ofmulathenumericalapproximationfaclassofstochasticdelayneural

5、wkswithPoissonjumpisdiscussed.Asufficientconditionftheconvergenceofthenumericalapproximationsolutiontotheanalyticsolutionisgiveninmeansquareatlastthenumericalmethodisvalidatedthroughanumericalexample.3.Inthispaperapplyin

6、gIt?ofmulaBurkholderDavisGundyinequalityCauchyinequalitytheGronwallinequalityweprovedthattheEulerapproximatesolutionswillconvergetotheexactsolutionsfstochasticdelayneuralwkswithPoissonjumpMarkovianswitchingunderlocalLips