PI Lecture Series: Jörg Raisch: Models for Feedback – Costs and Benefits of Simplicity

PI Lecture Series

Models for Feedback — Cost and Benefits of Simplicity

Joerg Raisch, Control Systems Group, TU Berlin

Abstract:

We’ll start with a short introduction to feedback control. I’ll
emphasise that the systematic design of feedback requires a formal
model of the system to be controlled, and that the choice of such a
model represents a fundamental degree of freedom in the design
process. In general, one is interested in “simple” models, where the
notion of “simplicity” may refer to the dimension or the structure
(e.g., linearity) of the model. We’ll explore possible costs that are
related to model simplification: examples are the loss of predictive
power or a weaker algebraic setting. We will finally discuss
conditions that make these costs acceptable.