A new approach to stability analysis of discrete-time recurrent neural networks with time-varying delay

In this paper, the problem of stability analysis of discrete-time recurrent neural networks with time-varying delay is studied. Based on the general assumption of time delay (that is 0 < dm ≤ d(k) ≤ dM), we represent d(k) as dm+h(k) with 0 ≤ h(k) ≤ dM-dm, and introduce a new Lyapunov functional with the idea of delay partitioning. A new stability criterion is then obtained by utilizing the most updated techniques for achieving delay dependence, which is characterized in terms of linear matrix inequalities (LMIs) and can be easily checked by utilizing the efficient LMI toolbox. The merit of the proposed stability lies in its less conservatism than most of the existing results, which is well illustrated via an example.