一类忆阻递归神经网络的全局指数稳定性

Global exponential stability of a class of memristor-based recurrent neural networks

研究一类忆阻递归神经网络(MRNNs)平衡点的全局指数稳定性。MRNNs激励函数是Lipschitz连续的,利用线性矩阵不等式和Lyapunov泛函理论,证明MRNNs平衡点是全局指数稳定的,并且在未引入其他参数的情况下,得到MRNNs平衡点全局指数稳定的充分条件,为电路的设计与实现提供了保障。数值实例验证了结果的有效性。

The global exponential stability of the equilibrium point of a class of memristor-based neural network （MRNNs） is studied. The activation function of the MRNNs is Lipschitz continuous, by using linear matrix inequality and Lyapunov functional theory, it is proved that the equilibrium of MRNNs is globally exponentially stable. Sufficient conditions are obtained for global exponential stability of the equilibrium point of MRNNs without other parameters, so as to provide a guarantee for designing and realization of the circuit. Numerical examples are illustrated to verify the effectiveness of the proposed results.