Distributed MPC for tracking based on reference trajectories

In this paper, we consider the control of large-scale processes with both input and state couplings. A distributed model predictive control (MPC) strategy for tracking based on the reference trajectories is presented. The proposed distributed MPC strategy requires decomposing a large-scale system into several smaller ones and solving convex optimization problems independently. Distributed MPC tracking strategies for unconstrained and constrained processes are presented, respectively. An iterative algorithm is presented to coordinate the distributed MPC controllers. The proposed algorithm is applied to a four-tank process to demonstrate the effectiveness.