Some results on the stability of analog neural networks with time delays

This paper investigates the global asymptotic stability (GAS) independent of time delay for analog neural networks models whose self-feedback interconnections have and have no corresponding time delays respectively. By constructing two different Lyapunov functionals, some new sufficient conditions represented by M-matrices, which can guarantee the GAS independent of time delay for the analog neural network models with time delays, are derived. These sufficient conditions obtained for GAS are also compared with the existing ones in the literature.