Adaptive RL Optimized Bipartite Consensus Tracking for Heterogeneous Nonlinear MASs Under a Switching Threshold Event Triggered Strategy

Combining with the reinforcement learning (RL) strategy and adaptive backstepping method, an event-triggered-based adaptive optimized bipartite consensus control protocol is constructed for heterogeneous nonlinear strict-feedback multi-agent systems (MASs). The control objective is to synchronize the state variables of the MASs and obtain the optimized bipartite tracking performance. In order to finish the control objective, a fuzzy logic system (FLS) based RL method is employed in the critic-actor structure. Different with the exist RL methods, the RL updating laws in this article is stemmed from a simple positive function which is derived by Hamilton-Jacobi-Bellman (HJB) equation. The FLSs are used to release the persistence excitation condition and the switching threshold event-triggered strategy is designed to reduce the calculation and communication frequency. It is shown that the developed control protocol not only achieves the bipartite consensus tracking performance of the studied heterogeneous MASs but also maintains the stability of the whole closed-loop system. Finally, the simulation results are provided to demonstrate the effectiveness of proposed control protocol.
Note to Practitioners
—Nowadays, energy conservation and environmental protection have become the current trend. Therefor, it is imperative to consider efficiency improvement as the fundamental principle of modern control design. In this paper, an optimal control method is employed to find a control protocol for considered MASs to accomplish the control task with the minimum utilization of control resources. In some practical application scenarios such as sea voyage for deep-sea exploitation or ocean transportation, the designed optimal control protocol can not only reduce the massive energy consumption for surface vessels but also eliminate the persistence of excitation assumption in Zhang and Feng (2021), Tang and Lau (2022), and Ren et al. (2022). The developed switching threshold event-triggered strategy and bipartite control design can effectively reduce the call frequency of channel resources in practical and divide all agents into two groups in surface vessels collaboration. Thus, our control algorithm can be easily applied in practical engineering.