Robust coalitional distributed model predictive control algorithm with stability via terminal constraint

This paper proposes a robust coalitional distributed model predictive control algorithm suitable for input coupled sub-systems. The core idea is based on a distributed iterative implementation with minimum information exchange, in which the coupling information is regarded as bounded additive uncertainty. If the disturbance received from the neighbours is too large for the local optimization problem, then a coalitional strategy is adopted. The closed-loop stability is guaranteed via terminal constraint. The performance of the coalitional strategy is compared with an iterative min-max distributed model predictive controller and a centralized model predictive controller.