Global Stability Criterion of Memristor-Based Recurrent Neural Networks with Time Delays

The paper investigates the uniform asymptotic stability of memristor-based recurrent neural network with time delays. Uniqueness of the equilibrium point of memristor-based neural networks is proved by constructing the Lyapunov energy function, employing homeomorphism mapping principle and differential inclusion. Sufficiency criterion based on Mmatrix is proposed to confirm the equilibrium is global asymptotic stable. The deduced criterion extends the result based on M-matrix, which has certain robustness for different time delays and activation functions. According to the physical parameters of the system. Numerical analysis and simulation results are presented to demonstrate effectiveness of the criterion.