Dynamic Event-Triggered DMPC With Variable Prediction Horizon for Disturbed Nonlinear Multi-Agent Systems

This article investigates the formation stabilization problem of continuous-time nonlinear multi-agent systems subject to state constraints, input constraints, and external disturbances. To solve this issue, a dynamic event-triggered distributed model predictive control algorithm is developed, integrating a control configuration that simultaneously considers both the triggering scheme and the variable prediction horizon. Specifically, a dynamic event-triggered mechanism based on feasibility analysis is proposed to adaptively adjust the triggering threshold, thereby reducing computational and communication burdens while preventing Zeno behavior. Meanwhile, a variable prediction horizon scheme is designed for each agent to effectively shorten the prediction horizon of the involved optimal control problem, which reduces the computational complexity of the proposed algorithm. Furthermore, theoretical conditions are established to ensure the recursive feasibility and closed-loop stability of the algorithm. Finally, theoretical results are verified through a numerical example with comparison analysis.