Round-robin scheduling protocol-based stabilization for discrete-time Markov jumping neural networks

Abstract

This paper focuses on the stabilization problem of discrete-time Markov jumping neural networks (DMJNNs) under communication constraints between the actuators and the remote controller. To reduce the communication burden, a round-robin scheduling protocol is introduced to allocate the right to a single actuator to exclusively use the communication channel at a given time while ensuring that multiple nodes can take turns to occupy the channel for signal transmission. Two criteria for stochastic analysis of the closed-loop network model are presented based on different assumptions of the activation functions. Then, numerically tractable design methods for the required controller gains are developed. Finally, the three-mode DMJNN model is used as an example to illustrate the complementarity and effectiveness of the presented analysis results and design methods.