Integral reinforcement learning-based online adaptive event-triggered control for non-zero-sum games of partially unknown nonlinear systems

This paper develops an integral reinforcement learning (IRL)-based adaptive control method for the multi-player non-zero-sum (NZS) games of the nonlinear continuous-time systems with partially unknown dynamics, in the context of event-triggered mechanism. With the principle of IRL method, the requirement for the system drift dynamics is relaxed in the controller design. Moreover, different from the conventional iteration computation methods, the algorithm developed in this work is implemented in an online adaptive fashion, which provides a new way to combine the IRL algorithm and the event-triggered control framework in solving the NZS game issues. In the event-based algorithm, a state-dependent triggering condition is presented, which not only guarantees the closed-loop system stability, but also reduces the computation and communication loads of the controlled plant. By means of Lyapunov theorem, the uniform ultimate boundedness (UUB) properties of the system states and the critic weight estimation errors have been proved. Finally, two numerical examples are utilized to demonstrate the efficacy of the proposed method. (c) 2019 Elsevier B.V. All rights reserved.