New conditions for exponential stability of delay impulsive neural networks

Abstract

Impulsive effects, which widely exist in various dynamical systems, including neural networks, can influence the dynamic behavior of systems just as delayed effects. A generalized model of neural networks involving variable delays and impulses is formulated. By introducing differential inequality with impulsive initial conditions and employing the properties of the M-matrix, we obtain new sufficient conditions ensuring global exponential stability of the impulsive delayed system. The results extend and improve those of earlier publications. An example and simulation are given to illustrate the theoretical results.