Stability analysis of discrete-time delayed neural networks via delay-product-type Lyapunov-Krasovskii functionals

This article is concerned with the stability problem for discrete-time neural networks with a time-varying delay. For the two cases that the delay-variation bounds are known and unknown, new augmented Lyapunov-Krasovskii functionals (LKFs) are correspondingly constructed by fully considering the information on the state-related vectors and nonlinear activation function. Through the entire vector-extension method, the forward differences of the new LKFs are estimated to be affine with the delay. Relaxed stability criteria are consequently derived via the convex method. Two numerical examples are provided to show the effectiveness of the proposed method on conservatism reduction.