Stochastic Predictive Control: Theory, Algorithms, and Applications

This workshop has been cancelled!

Duration:

Sunday June 5th 13.30-17:30, lunch from 12:30-13:30.

Organizers and Speakers:

Ali Mesbah (University of California-Berkeley, USA) Victor M. Zavala (University of Wisconsin-Madison, USA)

Overview:

Systematic consideration of uncertainties in optimal control is a challenging problem that has been the subject of extensive research. This workshop will present advances and challenges of different approaches for stochastic predictive control. In particular, we will discuss advantages and limitations of sample-based and polynomial chaos frameworks for general classes of complex systems and different formulations that include risk functions and probabilistic constraints. After laying the theoretical foundation for these frameworks, we will discuss how uncertainty modeling plays a key role in choosing the particular framework. We will demonstrate the concepts using several real-world biological, chemical, and energy applications. Finally, we will present existing software tools and discuss open challenges in the area. The workshop is designed for researchers with a basic knowledge of statistics and model predictive control, who intend to gain an overview of the implications and challenges arising in the control of complex systems with explicit handling of system uncertainties.

Outline:

This half-day workshop will consist of two parts. The first part will provide a concise overview of theory and formulations for stochastic model predictive control and introduce the polynomial chaos framework. In the second part, we will discuss the sample-based framework and will introduce software tools for formulating stochastic MPC problems. We will demonstrate the software tools by several applications to complex biological and energy systems. All workshop notes and slides will be provided, and computer codes for selected examples and applications will be made available on the workshop's website. The outline of the workshop is as follows:

• Opening remarks (15 min)
• Introduction to Stochastic MPC formulations (45 min)
• The generalized polynomial chaos framework (45 min)
• The Sample-Based Framework (45 min)
• Applications and Software Tools (45 min)
• Wrap-up (15 min)