Global Exponential Stability of Discrete-Time BAM Neural Networks With Variable Delays

Without the boundedness assumption for the activation functions and without any extra constraints on variable delay functions, based on Lyapunov functional, homeomorphism map, techniques of inequalities and linear matrix inequality(LMI), a new delay-dependent stability criterion for (i) existence (ii) uniqueness and (iii) globally exponential stability of equilibrium point, of a class of discrete-time bidirectional associative memory (BAM) neural networks with variable delays is derived.