On distributed model predictive control for systems with decoupled dynamics

This paper studies distributed model predictive control for the system which is composed of several subsystems with decoupled dynamics. Couplings between different subsystems are manifested by the cost function and constraints. The couplings are described through a connected graph. In this connected graph, each code denotes a subsystem. The cost function and constraints of a code are the functions of its state and the states of its neighbors. The centralize MPC controller for the over system is broken into distinct distributed MPC controllers of smaller size. Each MPC controller is associated to a different node. The distributed MPC and centralize MPC are compared through a numerical example.