Convergence properties of two coordinated distributed MPC algorithms

Computational properties of distributed model predictive control (DMPC) are seldom studied in the literature. In this work, we focus on the convergence properties of two co-ordinated DMPC (CDMPC) algorithms: the prediction-driven CDMPC and the price-driven CDMPC. By restricting this study to linear unconstrained systems, the two CDMPC algorithms are first transformed into iterative forms. Subsequently, convergence conditions and rates for the two algorithms are derived. The applicability of the theoretical results is illustrated via extensive numerical experiments.