Impulsive Stabilization of Unconstrained Multilayer Recurrent Neural Networks with Node-Based Time-varying Delays

Abstract

This article discusses the exponential stabilization of node-dependent delayed multilayer neural networks (NDDMNNs) under impulsive control. To address different modeling requirements in complicated applications, node-based interlayer and intralayer parameters are presented to design the neural network model, indicating that The nodes constituting the network can have different structures. Meanwhile, the novel model considers the node-dependent time-varying delays, and this article develops the sparse matrix approach to translate the node-dependent delayed NDDMNNs model into an multiple delayed model, ensuring that the vector form of NDDMNNs can be constructed and studied by using existing technical approaches. Then, an analytical framework with super-Laplacian matrix and time-dependent Lyapunov function methods is proposed to derive exponential stabilization results. Finally, a numerical simulation example is given to verify the obtained results.