Bounded real lemmas for inertial neural networks with unbounded mixed delays and state-dependent switching.

This paper mainly studies the bounded real lemma for inertial neural networks with unbounded time-varying transmission delays, unbounded distribution delays and state-dependent switching. Bounded real lemmas of inertial neural networks under consideration are presented by proposing a parameterizing approach based on the system solutions. The advantage of this approach is that it neither decomposes the model into two first-order differential equations nor constructs any Lyapunov–Krasovskii functional, thus reducing computational effort and complexity. Furthermore, the obtained sufficient condition contains only a few simple linear scalar inequalities, which can be easily solved by using MATLAB. Finally, a numerical example and its numerical simulation are used to demonstrate the validity of the theoretical results. • A novel approach based on the system solutions is proposed for the first time. • The proposed method reduces computational effort and complexity. • The obtained sufficient condition can be easily solved by using MATLAB. [ABSTRACT FROM AUTHOR]