Two-layer Hierarchical Predictive Control via Negotiation of Active Constraints1

We study a constraint negotiation approach for coordination of distributed model predictive controllers. The constraint negotiation techniques represent feasible-path coordination mechanisms having a goal of achieving an optimal distribution of shared resources, that represent physical couplings, among the subsystems. We present recent developments of the coordination scheme that enhance its convergence.We apply the presented algorithm on a benchmark problem of hierarchical MPC to illustrate the effectiveness of the proposed coordination scheme.