Stability of almost periodic solution for a generalized neutral-type neural networks with delays

In this paper, a generalized neutral-type neural networks with delays is studied. Some simple sufficient conditions are obtained for guaranteeing the existence, global asymptotic stability and exponential stability of the unique almost periodic solution for the system by using fixed point theorem, Lyapunov functional method and comparison theorem, respectively. It is interesting that, in some special cases, we show that the condition for guaranteeing the existence of almost periodic solution also yields the global exponential stability of it.