KU Leuven, Belgien
Monday, February 08, 2016, 11:00
Room 01-012, Georges-Köhler-Allee 102, Freiburg 79110, Germany
While many engineering problems are nowadays translated into a numerical optimization problem, the associated numerical data are often inaccurate or uncertain. As this uncertainty of data may turn the numerical optimum into a poor or even unacceptable solution for the true problem, there is a great push for robust optimization techniques suited for engineering applications.
I will present a novel relaxation scheme for robust optimization problems that outperforms existing schemes in generality and numerical efficiency. In addition, I will extend the scope of robust optimization beyond its original raison d’être by demonstrating its potential in applications that are free of uncertainty.