Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Robust convex optimization and its application to advanced control engineering

The goal of this presentation is to show how emerging techniques from the new mathematical field of robust convex optimization can be employed to provide robust solutions to validation, analysis and worst case simulation of linear models associated with dynamical systems affected by lp-norm bounded uncertainty. The key-role to the success of our method [1] is represented by the ability one has to determine the maximal level of uncertainty for which structural LMI-certifiable uncertain system properties admit a robust property certificate which is the same with determining the supreme of all uncertainties levels for which a semi-infinite semi-definite system of matrix inequalities admits a common solution. While the latter problem is NP-hard, we shall show that under realistic circumstances one can determine a robust certificate and a tight approximation of the robustness margin. The technique we use is a lift-and-project scheme which was developed by Ben-Tal & Nemirovski [2] in order to solve the so-called Matrix-Cube Problem. Approximate solutions to the problems of validation, analysis and worst case simulation of linear models are developed accordingly within this framework.
Type of Seminar:
Public Seminar
Dr. Florin Barb
Division Control & Simulation, Faculty of Aerospace Engineering, Delft University of Technology, Netherlands
Jun 16, 2003   16:15

ETH Zentrum, Maschinenlabor, Sonneggstr.3, ML E 12
Contact Person:

Prof. L. Guzzella
No downloadable files available.
Biographical Sketch: