Recently we have extended the model predictive control (MPC) framework to some classes of hybrid systems such as traffic-signal-controlled intersections, and first order linear hybrid systems with saturation. General hybrid systems arise from the interaction between continuous variable systems (i.e., systems that can be described by a difference or differential equation) and discrete event systems (i.e., systems where the state transitions are initiated by events that occur at discrete time instants). We could say that a hybrid system can be in one of several "modes" whereby in each mode the behavior of the system can be described by a system of difference or differential equations, and that the system switches from one mode to another due to the occurrence of events (see the figure below). Typical examples of hybrid systems are flexible manufacturing systems, transportation networks, beer brewing processes, logistic systems, etc.

A schematic representation of a hybrid system with two regimes.

The aim of this proposal is to further investigate several aspects of MPC for hybrid systems: for which classes of systems can the resulting optimization problems be solved efficiently (e.g., via convex optimization), development of efficient algorithms, stability issues, inclusion of disturbances, noise and modeling errors, implementation issues (prediction, partial information, when to update), and actual implementation on a practical example.

If you are interested in selecting this project as your MSc project, please come along or send us an email for more information.

This page is maintained by Bart De Schutter.