Passivity analysis for uncertain BAM inertial neural networks with time-varying delays

Without reducing the second order terms into first order, this paper analyzes the passivity for uncertain BAM inertial neural networks with time-varying delays. Two different types of uncertainties covering parameter and Markovian Jump uncertainties are considered, respectively. Based on Lyapunov functional method, inequality techniques and applying relaxed integral inequality, a few new Lyapunov functionals are proposed to straightway address the passivity of the concerned systems, several delay-dependent criteria are acquired in terms of linear matrix inequalities to insure the passivity. Moreover, the time varying delays here can be non-differentiable and the conservatism of the obtained results is further reduced by using relaxed integral inequality. Finally, some numerical simulations are provided to demonstrate the feasibility of proposed results. (c) 2020 Elsevier B.V. All rights reserved. Without reducing the second order terms into first order, this paper analyzes the passivity for uncertain BAM inertial neural networks with time-varying delays. Two different types of uncertainties covering parameter and Markovian Jump uncertainties are considered, respectively. Based on Lyapunov functional method, inequality techniques and applying relaxed integral inequality, a few new Lyapunov functionals are proposed to straightway address the passivity of the concerned systems, several delay-dependent criteria are acquired in terms of linear matrix inequalities to insure the passivity. Moreover, the timevarying delays here can be non-differentiable and the conservatism of the obtained results is further reduced by using relaxed integral inequality. Finally, some numerical simulations are provided to demonstrate the feasibility of proposed results.

Automated summary

This paper investigates the passivity of uncertain BAM inertial neural networks with time-varying delays, proposing new Lyapunov functionals and delay-dependent criteria based on linear matrix inequalities to ensure the passivity of the systems. Numerical simulations demonstrate the effectiveness of the proposed approach.