Distributed Model Predictive Control: Two-Controller Cooperation

Abstract

In Chap. 4, a class of distributed control problems is studied. This class of distributed control problems may arise when new control systems which may use networked sensors and actuators are added to already operating control loops designed via model predictive control (MPC) to improve closed-loop performance. To address this control problem, a distributed model predictive control method is introduced where the preexisting control system and the new control system are redesigned/designed via Lyapunov-based MPC. The distributed control design stabilizes the closed-loop system, improves the closed-loop performance and allows handling input constraints. Furthermore, the distributed control design requires that these controllers communicate only once at each sampling time and is computationally more efficient compared to the corresponding centralized model predictive control design. The distributed control method is extended to include nonlinear systems subject to asynchronous and delayed measurements. Using a nonlinear reactor–separator example, the stability, performance and robustness of the distributed predictive control designs are illustrated.