Almost periodic solution of Cohen–Grossberg neural networks with bounded and unbounded delays

In this paper, a class of Cohen–Grossberg neural networks with bounded and unbounded delays is discussed. Several new sufficient conditions are obtained ensuring the existence and exponential stability of the almost periodic solution for this model based on inequality analysis technique and combing the exponential dichotomy with fixed point theorem. The obtained results are helpful to design globally exponentially stable almost periodic oscillatory neural networks. Two numerical examples and simulations are also given to show the feasibility of our results.