Model-Free Event-Triggered Optimal Containment Control for Multiagent Systems via Adaptive Dynamic Programming

In this article, a novel model-free event-triggered method considering the distributed quadratic index is proposed for the containment control problem of discrete-time linear multiagent systems (MASs). To save computational resources and improve control efficiency, a triggering threshold function is designed for each agent based on the distributed performance function. The control protocol of each agent is updated only when the Euclidean norm of the triggering error exceeds the threshold. In order to obtain the model-free event-triggered control solution, an online actor–critic framework is established to approximate the event-triggered control protocol and performance function of each agent. It is proved that the containment error of MASs is asymptotically stable under the action of the designed event-triggered condition, and the weight estimation errors of actor–critic neural networks are uniformly ultimately bounded. Finally, numerical simulation results are given to verify the effectiveness of the proposed method.