The Study Of Bipartite Synchronization For Coupled Memristive Neural Network With Time-Varying Delay

Compared with the traditional neural networks,the memristive neural networks using memristors instead of ordinary resistors has self-learning ability,associative storage function and adaptability,which is closer to the model of human brain.In real life,interpersonal relationships often include cooperation and competition.Similarly,there will also be competition and cooperation among each neuron node in the coupled memristive neural networks.In this paper,the bipartite synchronization problem of coupled memristive neural networks with uncertain disturbances and time-varying delays are considered,and the sufficient conditions for the network to achieve finite time bipartite synchronization are obtained.The specific works are as follows:(1)The problem of finite time bipartite synchronization with the same lag in a cooperative-competitive network is studied.The coupled memristive neural network is transformed into a continuous system by utilizing set-valued mapping and differential inclusion theory.And a two-phase control strategy is designed.Taking synchronization error as the switching index,in the first phase,when the switching index is greater than the given switching value,a impulsive control is used to make the error of the coupled memristive neural network and its target network exponential converge to the interior of the sphere with the switching value as the radius;in the second phase,when the control index is less than the given switching value,a feedback control with a sign function is used to make the network error converge to zero in a finite time.The synchronization criteria for coupled memristive neural networks with time-varying delays and uncertain disturbances are obtained.Compared with existing results,the restriction of the switching value on controller in this paper is relaxed.(2)The problem of finite time bipartite synchronization with different lags in cooperative-competitive network is studied.According to the differences of competitive-cooperative relationship,different forms are used to express coupling terms in the network model.And a hybrid control protocol that can handle different transmission lags between networks is proposed.Using the Lyapunov function method and matrix inequality analysis techniques,the conditions of the finite time bipartite synchronization with different delays for coupled memristive neural networks with uncertain disturbances and time-varying delays are proposed.(3)The finite time bipartite synchronization problem of fuzzy coupled memristive neural networks is studied.Based on the T-S fuzzy model,a fuzzy coupled memristive neural network is obtained by combining the coupled memristive neural network with fuzzy rules.In order to deal with attacks on the communication channel from the controller to the actuator,a two-phase controller that is subject to deception attack and DoS attack respectively is designed.The sufficient conditions for achieving finite time bipartite synchronization of fuzzy coupled memristive neural networks with uncertain disturbances and multiple delays are derived,and the setting time to achieve synchronization is given.