New asymptotic stability analysis for generalized neural networks with additive time-varying delays and general activation function

Abstract

This paper focuses on the asymptotic stability problems of the generalized neural networks (GNNs) with two additive time-varying delays (ATDs) and general activation function. First of all, a simple Lyapunov-Krasovskii functional (LKF) is established. Then not only the generalized free-weighting-matrix-based (GFWM-based) inequality together with its optimization strategy on choosing an arbitrary vector is utilized to estimate the single integral terms in derivative of the constructed LKF, but also the general activation function method is applied to introduce more information on cross terms of neuron activation function. Finally, a less conservative stability criterion together with its corollary is derived with convex combination technique, whose feasibility and superiority can be demonstrated with two numerical examples.