求解微波電磁場(chǎng)反問(wèn)題的神經(jīng)網(wǎng)絡(luò)方法.pdf

1、大連理工大學(xué)碩士學(xué)位論文求解微波電磁場(chǎng)反問(wèn)題的神經(jīng)網(wǎng)絡(luò)方法姓名：蔡良弟申請(qǐng)學(xué)位級(jí)別：碩士專(zhuān)業(yè)：信號(hào)與信息處理指導(dǎo)教師：郭成安20050301求解微波電磁場(chǎng)反問(wèn)題的神經(jīng)網(wǎng)絡(luò)方法SolvingtheInverseProblemofMicrowaveFieldBasedontheArtificialNeuralNetworkAbstractForthesakeoflargecomputationandgreatcomplexityofusin

2、gconventionalapproachtosolvetheinverseproblemofmicrowaveelectromagneticfieldinthispaperwewillapplyartificialneuralnetworkt0solvethiskind0finverseproblemDuetothenonlinearnatureoftheartificialandparallellreatmentondatather

3、esolutionofthemicrowaveelectromagneticfieldinverseproblemCallbemadewithoutcalculationmodelsbutonlywiththesimplecoupledataabouttheirmicrowavefieldsystemthemodelCallbesimulatedbythestirstudyofthenetworkMoreoverthankstother

4、obustcharacterandresistantbotheredcapacityofthenetworktheresultwillnotbebotheredeveniftherearesomemistakeshdataTheabovefeathersoftheartificialneuralnetworkwillofferabroadfutureintheapplicationofthemicrowaveeleclromagneti

5、cfieldinverseproblemToobtaintheaccuratedatawhichitisusedtotraintheartmcialn涮networkweapplytheanalyticalsolutionapproachbasing012theoperation’sEigenfunctiontosolvethecomputationofthemicrowaveelectromagneticDuringthetraini

6、ngprocess，wewillhavetheConjugategradienttotrainartificialneuralnetworkandselectthelearningratebyourwaytoensurethevalidationoftrainingalgorithmInthepaperwewillcheckuptheinverseproblemsolutionwhichisobtainedbythewayofartif

7、icialneuralnetworkbysomenumericalexperimentbytheexperimentresultwecanincludethatsolvingtheinverseproblemofmicrowaveproblembasing01“1theartificialneuralnetworkiSmoreaccurate也anthetraditionalapproachKeyWords：TheInverseprob