Synchronization Analysis Of Cohen-Grossberg Memristive Neural Networks

Neural networks are complex dynamic models used to simulate the human brain,which can be applied to many natural disciplines,and neural networks have rich dynamic properties in the application process,and this paper mainly analyzes their synchronization.In addition,for the Cohen-Grossberg neural network model with memristoria,the addition of memristor is more effective than the ordinary neural network for simulating the brain,and the universality of the Cohen-Grossberg neural network is stronger.On the other hand,in the discussion of the various dynamic properties of neural networks,time lag and external interference are inevitable,so time-delay neural networks with external interference also need to be considered.Based on the above considerations,this paper mainly studies the finite time synchronization and fixed time synchronization of memristem-Grossberg neural network with discontinuous activation function.The main research contents are as follows:First,the fixed time synchronization of memristor Cohen-Grossberg neural networks with time-varying delays and discontinuous activation functions is discussed.A new kind of sliding mode surface is established,and the Gudermannian function is used to guarantee the dynamics on the sliding mode surface in fixed time.Considering the influence of time delay,two different control schemes are introduced to ensure the fixed time reachable of the sliding mode.In addition,some sufficient criteria for fixed time synchronization of neural networks are given,and the settling time is estimated.Finally,numerical simulation is given.Secondly,the fixed time synchronization of Cohen-Grossberg memristor neural networks with mixed delay and discontinuous activation functions is studied.An appropriate controller is designed to achieve the system to reach the sliding mode surface in fixed time by using the existing theory of fixed time stability,and surface in fixed time.In addition,some sufficient criteria for fixed time synchronization of neural network are given,and the settling time is estimated.Finally,the design of a uniform framework for finite-time and fixed-time synchronization of Cohen-Grossberg memristor neural networks with delayed and discontinuous activation functions is discussed.By introducing a new uniform integral sliding mode manifold and giving the corresponding uniform controller,the criterion to solve the related problems is established,and the estimation of settling time is given.The proposed unified framework can obtain different control schemes by selecting different parameters of the controller and sliding mode manifold,which extends the previous results.Finally,the correctness and superiority of the conclusion are proved by some numerical examples.In this paper,the synchronicity analysis of Cohen-Grossberg memonic neural network is carried out,especially a general sliding mode control is used to synchronize the neural network in a limited time and a fixed time,which supplements the results of the study on the synchronicity of the neural network.