Coalitional MPC with predicted topology transitions

Abstract This paper proposes a novel hierarchical coalitional MPC technique, where transitions to the best communication topology are considered over the prediction horizon. For this reason, a new variable, called transition horizon, is added to the optimization problem to compute the optimal instant to introduce a new topology. Consequently, local controllers can anticipate topology transitions and adapt their trajectories whilst optimizing their local interests. Furthermore, stability guarantees in the closed-loop control of each coalition are provided. The benefits of this control method are shown via a simulated non-linear eight-coupled tanks plant.