Reinforcement learning-based fixed-time synchronization control of chaotic systems with event-triggering mechanism and its application
摘要
This paper investigates the optimal fixed-time synchronization control problem for chaotic systems via a backstepping-based framework. First, an actor-critic architecture is constructed within the backstepping procedure to design the optimal fixed-time controller, and neural networks are employed to compensate for unknown system nonlinearities, thereby achieving synchronization of the master-slave systems. Then, a dynamic event-triggering mechanism is incorporated to effectively reduce the controller execution frequency. Moreover, it is worth emphasizing that the proposed backstepping-based optimal control scheme ensures that the synchronization errors converge to a bounded region within a fixed time interval and effectively excludes Zeno behavior. Finally, numerical examples validate the effectiveness of the proposed optimal fixed-time control approach and apply it to image encryption and decryption.