Jump to Main Content
A Bilevel Optimization Approach to Coordination of Distributed Model Predictive Control Systems
- Hassanzadeh, Bardia, Liu, Jinfeng, Forbes, J. Fraser
- Industrial & engineering chemistry process design and development 2018 v.57 no.5 pp. 1516-1530
- algorithms, controllers, process design
- A novel algorithm is proposed to coordinate distributed model predictive control (DMPC) systems. The main goal of coordinated distributed model predictive (CDMPC) is to improve the performance of an existing decentralized MPC system, which neglects the interactions among subsystems. In the first scenario, an analytic approach derives the global optimal solution to the plant-wide problem in the absence of inequality constraints in local controllers. In the second scenario, an iterative approach determines a local optimal solution to the general CDMPC problem based on the method of feasible directions. The proposed CDMPC algorithm is capable of stabilizing open-loop unstable dynamics.