ME-425 Model Predictive Control

Instructor: Prof. Colin Jones, Laboratoire d’Automatique

Assistants: Jean-Hubert Hours, Andrea Alessandretti

This course will be offered for the first time in Spring, 2012.

The theory and practice of model predictive control has grown rapidly in the last two decades, with important advances in our theoretical understanding coming from academia, as well as a maturing of the technology within the petrochemical industry and the uptake as the preferred control technique in many novel areas. This success is a result of the main benefits of MPC: the very flexible specification of time-domain objectives, the performance optimization of highly complex multivariable systems and the ability to explicitly enforce constraints on system behavior. The key difference to other control methods is that an MPC strategy places an optimization routine in the control loop, which determines the control action to take by optimizing a performance index with respect to a prediction of future system behavior. This course will provide an introduction to the theory and practice of the most significant Model Predictive Control paradigms.

Course details and lecture slides are available on the EPFL moodle site.

If you are not from EPFL and would like access to the course material, please email Colin Jones, and I’ll be happy to send you everything.