Integral Reinforcement Learning Control for a Class of Unknown Nonlinear Systems with an Application to a Microgrid System
Shanyong Xu, Hanguang Su, Xiaodong Liang, Jinzhu Yang, Jiawei Wang, Xinyang Luan · 2023
In this paper, a brand-new technique based on integral reinforcement learning (IRL) combined with the event-triggered control (ETC) for multiplayer non-zero-sum (NZS) game is proposed, taking into account nonlinear systems with uncertain system drift dynamics. System drift dynamics are no longer necessary for controller design with the IRL method. Furthermore, this method is implemented online, in contrast to other iterative calculating techniques. In this instance, the NZS game problems can be resolved by combining the IRL algorithm and the event-triggered control architecture. It offers a new triggering condition and lessens the computational and communication overhead of the entire control process. The system’s stability is ensured at the same time. An example is then given to show how well our method works.