The Stability Analysis Of A Class Of Delayed Recurrent Neural Networks

This paper is concerned with the problem of asymptotic stability for a class of recurrent neural network with time-varying delay.The time-delay in a system has great influence on its dynamic characteristics,which often leads to instabil-ity of the system and may even produce chaos phenomenon or periodic oscilla-tion.Therefore,the research on the delayed neural network is more complex and more practically valuable.In this paper,different methods are used to study the asymptotic stability of recurrent neural network with time-varying delay.(1)The problem of stability analysis of a class of delayed recurrent neural network with time-varying delay is concerned,and the neuron activation func-tion is assumed to satisfy a general sector condition.By using the Wirtinger-based inequality and reciprocally convex approach to estimate the derivative of Lyapunov-Krasovsii functional,a new delay-dependent stability criterion is ob-tained.In addition,the convex hull technique is applied to deal with the deriva-tive of time-varying delay,the proposed criterion relaxes the restriction on the derivative of time-varying delay.A numerical simulation is given to verify the effectiveness of the proposed criterion.(2)The problem of asymptotic stability of a class of delayed recurrent neural network with time-varying delay is concerned.The activation function is mono-tone nondecreasing with known lower and upper bounds,and satisfies a more general sector condition.A novel Lyapunov-Krasovsii functional based on delay-decomposition approach is constructed.By using Jensen inequality and recipro-cally convex combination to estimate the derivative of Lyapunov-Krasovsii func-tional,a new stability criterion for recurrent neural network with time-varying delay is obtained.And a numerical simulation is given to verify the effectiveness of the proposed method.