Optimal self-triggered control and network co-design for networked multi-agent system via adaptive dynamic programming

Abstract

In this paper, optimal self-triggered control and network co-design problem has been studied for networked multi-agent system (MAS) with uncertain physical and network dynamics. There are two main challenges, i.e. 1) effective modeling of control and network jointly is lacking, 2) uncertain physical and network dynamics complicate the optimal design significantly. To overcome these challenges, a novel control and network joint modelling algorithm has been developed for networked MAS firstly. Specially, we consider the self-triggered control design from physical system aspect and admission control design from network aspect. Next, adopting adaptive dynamic programming strategy, a novel online time-based optimal self-triggered control and network admission co-design has been proposed. The developed scheme could not only optimize both self-triggering time and number of admitted agents, but also relaxes the requirement of networked MAS physical system and network dynamics. Eventually, simulation results are also included to demonstrate the effectiveness of proposed scheme.