Quantcast

An Improved Result on Dissipativity and Passivity Analysis of Markovian Jump Stochastic Neural Networks With Two Delay Components.

Research paper by Gnaneswaran G Nagamani, Thirunavukkarasu T Radhika, Quanxin Q Zhu

Indexed on: 16 Oct '16Published on: 16 Oct '16Published in: IEEE transactions on neural networks and learning systems



Abstract

In this paper, we investigate the dissipativity and passivity of Markovian jump stochastic neural networks involving two additive time-varying delays. Using a Lyapunov-Krasovskii functional with triple and quadruple integral terms, we obtain delay-dependent passivity and dissipativity criteria for the system. Using a generalized Finsler lemma (GFL), a set of slack variables with special structure are introduced to reduce design conservatism. The dissipativity and passivity criteria depend on the upper bounds of the discrete time-varying delay and its derivative are given in terms of linear matrix inequalities, which can be efficiently solved through the standard numerical software. Finally, our illustrative examples show that the proposed method performs well and is successful in problems where existing methods fail.