Abstract:

Automatic control has a long history, beginning well before the age of digital computers. However, many of the modern methods involve optimization, and the more successful control methods involve real-time solution of mathematical programs. This presentation will begin with a brief introduction to the formulation (Model Predictive Control) using mathematical programming. One of the limitations of this method is its assumption that the dynamic model is without error. The major topic of the research involves extensions to model predictive control that explicitly account for uncertainty. The general approach using stochastic optimization principles leads to formulations that are intractable for real-time application, so modifications that lead to tractable, �good� solutions will be presented. The presentation will conclude with an application of the method to an industrial supply chain problem that requires frequent re-optimization due to model uncertainty.