GLOBAL EXPONENTIAL STABILITY OF COHEN-GROSSBERG NEURAL NETWORKS WITH TIME-VARYING DELAYS

Abstract

In this paper, without the assumptions for boundedness, monotonicity, and differentiability on activation functions and symmetry of interconnections, a class of Cohen-Grossberg neural networks with time-varying delays is studied.  A new useful criteria on the uniqueness of equilibrium is obtained by utilizing the nonlinear measure.  Combining with Dini derivatives and Young inequality, new sufficient condition for the global exponential stability is established by directly estimating the upper bound of solutions of the system.  All results are presented in M-matrix form, which extended and generalized the corresponding results in previous literature.