Universität Wien
Achtung! Das Lehrangebot ist noch nicht vollständig und wird bis Semesterbeginn laufend ergänzt.

390003 UK VGSCO Distributionally Robust Optimization (2018S)

Prüfungsimmanente Lehrveranstaltung


Hinweis: Ihr Anmeldezeitpunkt innerhalb der Frist hat keine Auswirkungen auf die Platzvergabe (kein "first come, first served").


Sprache: Englisch



Block, March 13-23, 2018
Seminar Room 3.307 (3rd floor, Faculty of Business, Economics and Statistics)

Please note that there's a change of schedule for the first week:
Tue 13.02. 10:30 - 12:00
Wed 14.03. 14:30 - 16:00
Thu 15.03. 15:30 - 17:00
Fri 16.03. 9:30 - 12:30
Mo 19.03. 11:15 - 13:15 ACHTUNG TERMINÄNDERUNG!
Tue 20.03. 10:30 - 12:30
Wed 21.03. 15:00- 17:00
Thur 22.03 10:30 - 12:30
Fri 23.03. 9:30 - 12:30


Ziele, Inhalte und Methode der Lehrveranstaltung

Uncertainty is traditionally modeled via probability distributions. However, observable statistical data can often be explained by many strikingly different distributions. This "uncertainty about the uncertainty" poses a major challenge for optimization problems with uncertain parameters: estimation errors in the parameters' distribution are amplified through the optimization process and lead to biased (overly optimistic) optimization results as well as post-decision disappointment in out-of-sample tests. The emerging field of distributionally robust optimization (DRO) seeks new optimization models whose solutions are optimized against all distributions consistent with the given prior information. Recent research results have shown that many DRO models can be solved in polynomial time even when the corresponding stochastic models are intractable. DRO models also offer a more realistic account of uncertainty and mitigate the post-decision disappointment characteristic of stochastic models. The course will provide an overview of the state-of-the art in DRO, focusing mainly on the theory distributionally linear and convex optimization, data-driven distributionally robust optimization, as well as applications in finance, statistics and machine learning.

Art der Leistungskontrolle und erlaubte Hilfsmittel

There will be a 1h written exam in the last session. Students are allowed to bring along a handwritten cheat sheet (1 page A4, both sides).

Mindestanforderungen und Beurteilungsmaßstab

Good knowledge of Linear Algebra, Linear Programming and basic Probability Theory. Basic knowledge of Conjugate Duality and Convex Analysis.


Convex Optimization, Robust Optimization and Distributionally Robust Optimization.


Zuordnung im Vorlesungsverzeichnis

Letzte Änderung: Fr 31.08.2018 08:43