The Design of Coalitional Model Predictive Control for Large-Scale Systems

This study provides a hierarchical control approach for large-scale systems composed of interconnected subsystems. A local controller or agent that can communicate via a network manages each subsystem. The upper layer is responsible of network topology selection to determine the optimal trade-off between the communication load and the control performance. The selected network topology splits the set of agents into cooperating groups of agents referred to as coalitions. At the bottom layer, each coalition computes the value of its control input variable by employing a decentralized model predictive control (MPC) scheme. Recursive feasibility and closed-loop stability analysis are carried out by considering fixed and varied network topologies. As a final point, a numerical simulation is carried out to verify the performance of the control scheme that is proposed