Finite-time Stabilization For Two-classes Of Stochastic Neural Networks With Time-delay

The essence of the research of the neural networks is learning from the human brain.Artificial neural networks,as a very important technological means,plays an important role in many fields,and attracts many scholars's research.In this paper,the finite-time stabilization for two-classes of stochastic neural networks with time-delay is studied.?1?For a given energy-storing electrical circuit,taking the measurement error of the inductor L,capacitor C₁,C₂,···,C_N,resistance R₁,R₂,time-delay and stochastic dis-turbance into consideration,and establishing a stochastic interval system with time-delay and Markov switching.By the equivalent transformation of the matrix,a stochastic uncer-tain system with time-delay and Markov switching is obtained.The finite-time stabilization problem of stochastic closed-loop system by constructing Lyapunov-Krasovskii function-al is studied,combined with Schur complement lemma,definition of finite-time stability and linear?nonlinear?matrix inequalities.On the basis of the finite-time stabilization and the definition of passivity,we get the time-delay independent sufficient conditions of the finite-time passivity for the stochastic closed-loop system.In order to reduce the conservatism,we construct a new Lyapunov-Krasovskii func-tional,and obtain the time-delay dependent sufficient conditions for the finite-time passive of the stochastic closed-loop systems.Finally,the design methods of passive controllers with delay-independent and delay-dependent sufficient conditions are given in the form of linear inequalities.?2?For a class of inertial neural networks,which can be transformed into a first order differential system by the means of variable substitution.By constructing Lyapunov func-tions,combining the definition of finite-time stable in probability,Schur complementary lemma,linear matrix inequality technology and analysis method,we can obtain the suffi-cient conditions of the finite-time stabilization for the stochastic closed-loop system,and the design method of the feedback controller.On this basis,the finite-time stabilization for stochastic inertia neural networks with time-varying delay is studied.It is found that changing the design of the controller,the delay in the system can be counteracted.That is to say,the sufficient conditions of the finite-time stabilization for the closed-loop system with time-delay is the same as the sufficient conditions of the finite-time stabilization for the closed-loop system without time-delay.Finally,summarizing the work of the full text,and pointing out the research direction of the next step.In a word,the research on two-classes of stochastic neural networks with time-delay not only enriches the finite-time control theory of neural networks,but also extends the research method of finite-time control for neural networks.Numerical simulation examples also illustrate the correctness of the conclusion and the effectiveness of the method.