Explicit MPC in real-world applications: the PAROC framework
Additional Document Info
2016 American Automatic Control Council (AACC). PAROC is an integrated framework and software platform for the development of explicit MPC controllers for real-world applications. Starting from a validated mathematical representation of the system under consideration, suitable model reduction and system identification techniques provide an approximate model which is used in combination with state-of-the-art explicit MPC algorithms to solve the underlying optimal control problem offline. The resulting solution is then seamlessly incorporated into the high fidelity model based closed-loop system, where it is validated without the need to solve an optimization problem online. The capabilities and applicability of this framework are highlighted via two case studies: the control of a combined heat and power (CHP) cogeneration system for residential use, as well as the optimal control of intravenous anaesthesia. Challenges such as hybrid systems and nonlinearity are discussed, along with new directions in the development of powerful novel multi-parametric programming algorithms and tools for a wide range of optimization and control problems, thereby paving the way for the wider applicability of the proposed PAROC framework to real-world applications.