简介:ApplicationoftheRTNNmodelforasystemidentification,predictionandcontrol;AssociativeMemoryUsingRatioRuleforMulti-valuedPatternAssociation;Batch-to-BatchModel-basedIterativeOptimisationControlforaBatchPolymerisationReactor;BehaviouralPlasticityinAutonomousAgents:AComparisonbetweenTwoTypesofController;ChannelEqualizationUsingComplex-ValuedRecurrentNeuralNetworks;Classificationofnaturallanguagesentencesusingneuralnetworks;Combiningarecurrentneuralnetworkandtheoutputregulationtheoryfornon-linearadaptivecontrol。
简介:Arecurrentlocalmodelnetworkfor,nonlinearprocessmodeling.ARECURRENTNEURALNETWORKFORI-DPHASERETRIEVAL.ARecurrentNeuralNetworkforNonlinearConvexProgramming.ArecurrentneuralnetworkmodeloftheB.T.fed-batchfermentationprocess.ActiveControlofSoundbasedonDiagonalRecurrentNeuralNetwork.
简介:Adesignmethodfornoisecancellerusingrecurrentneuralfilter,ADRNN-baseddirectmulti-stepadaptivepredictorforintelligentsystems,Agenerationalterationmodelforevolvingrecurrentneuralnetworktopologiesalongwithweights,AHIGH-PERFORMANCEMULTI-PURPOSEDSPARCHITECTUREFORSIGNALPROCESSINGRESEARCH……
简介:基于线性时变系统的稳定性理论,李雅普诺夫直接法和Gerschgorin圆盘定理求得判定广义Lienard方程振动系统达到全局同步的几种不同的代数判据.理论上比较这些不同代数判据表明:根据李雅诺夫直接法得到的代数判据优于根据Gerschgorin圆盘定理得到的代数判据,而且通过适当选取李雅普诺夫函数可以得到更优化的代数判据.Rayleigh—Duffing方程作为数值算例进一步验证了理论结果.