A criterion for global asymptotic stability of Cohen-Grossberg neural networks with delays

In this paper, the Cohen-Grossberg neural network models with time delays are considered. By constructing an appropriate Lyapunov functional, sufficient criteria for global asymptotic stability of the network are derived. These criteria independent of the magnitudes of the delays are applicable for other Cohen-Grossberg neural network models. Our results are less conservative and restrictive than previously known results and can be easily verified. And the result has overcome the obvious drawback that previous works neglect the signs of the connecting weights, and thus, do not distinguish the differences between excitatory and inhibitory connections. It is believed that our results are significant and useful for the design and applications of the Cohen-Grossberg model.