Min-Max Coalitional Model Predictive Control Algorithm

This paper proposes a coalitional model predictive control algorithm with feasibility guarantees for systems with bounded additive uncertainties. This formulation, suitable for sub-systems coupled through the inputs, assumes the coupling variables as disturbances and ensures a robust consensus with minimum information exchange. The simulation results show that the coalitional method has similar behaviour to the fully centralized algorithm and improved performance with respect to the decentralized and the iterative min-max distributed model predictive controllers.