Stability analysis of dynamical neural networks with uncertain delays

The problem of stability of dynamical neural networks with uncertain delays is studied, where uncertain delays are assumed to be constant. In this paper, a new approach is developed to establish delay-dependent sufficient conditions for asymptotic stability of delayed neural networks. The new approach is a combination of the discretized Lyapunov functional method and the free-weighting matrix technique. The established delay-dependent sufficient conditions are expressed by means of linear matrix inequalities, and thus are easily checkable. The new delay-dependent stability conditions are further illustrated by numerical results and are also compared with the existing results.