Research On Stability Analysis Of Memristor-based Neural Networks

With the development of the theory of neural network,it has brought a great impact on human science and technology,and our natural cognition.Memristor is one of the four basic passive electronic components except the inductor,resistor and capacitor.Memristor is an electronic component that changes the resistance with the number of charges that flow through it and it is different from the ordinary resistor that it is nonlinear electrical characteristics.The number of charges can be stored in memristor,when the circuit is disconnected.The characteristics of this kind of features are very similar to the working characteristics of the synapses of our human beings.The study of the memristor and the neural network is of great significance,such as research and development has been the dream of the humanoid robot.This paper first introduces the concept and the significance of the research of neural network and memristor.The definition of stability of neural networks in dynamics and two kinds of neural network models are studied in this paper.We use the Lyapunov stability theory,the methods of M matrix and linear matrix inequality(LMI).The stability of two classes of memristor-based neural network model with time delay is studied.We also skillfully use techniques such as linear matrix inequality to determine the stability of the criterion,and the use of mathematical software to show its stability.