NOVEL STABILITY CRITERIONS FOR TWO TYPES OF RECURRENT NEURAL NETWORKS WITH TIME-VARYING DELAYS

Wenguang Luo, Yonghua Liu , Guangming Xie, and Hongli Lan

Keywords

Recurrent neural networks, stochastic fuzzy cellular neural networks,stability criterion, Lyapunov–Krasovskii functional, Linear MatrixInequality, time delay

Abstract

This paper discusses the stability problems about two types of recurrent neural networks (RNNs) with time-varying delays. By ap- plying the linear matrix inequality (LMI) and Lyapunov–Krasovskii functional (LKF), the novel sufficient condition of globally exponen- tial stability for RNNs with time-varying discrete and distributed delays is derived under the condition of the more general activation function. The new mean-square global asymptotic stability criterion for stochastic fuzzy cellular neural networks (SFCNNs) with time- varying delays is also derived by constructing a new LKF and using Ito’s stochastic stability theory. The obtained results are helpful to design the stability of above two systems. Finally, some illustrative examples are given to verify the feasibility and effectiveness of the results.

Important Links:



Go Back