Decisions under Uncertainty

Britta Peis & Catherine Cleophas, September 22 to 25, 2015


The goal of our course is to combine theory and praxis in the area of optimization under uncertainty. Methods used in order to take decisions, for example, in production or logistics, are usually based on certain knowledge of the input data. However, in real-life, these input data can vary; they might be highly uncertain.

Teaching Objectives

The goal of our course is to combine theory and praxis in the area of optimization under uncertainty. The participants of our course will learn about methods that have been developed and shown to perform well for optimization problems under uncertainty, both in theory and in application.


  • Master course "Operations Research 1“
  • and at least one further course of the Master ORM specialization


  • 70 % oral participation
  • 30 % colloquium



The workshop "Decisions under Uncertainty" takes place at Chez Philippe in Gemmenich (Belgium). The Venue is located at Rue des Ecoles 69, 4851 Gemmenich, Belgium (see here ).


Preliminary Program


09:15 - 09:30 Welcome Session
09:30 - 11:00 Lecture
11:30 - 13:00 Exercise Session
13:00 - 14:30 Lunch Break
14:30 - 16:00 Lecture
16:30 - 18:00 Exercise Session
18:00 Open Evening Program (including time for homework and self-organized dinner)


09:30 - 11:00 Lecture
11:30 - 13:00 Exercise Session
13:00 - 14:30 Lunch Break
14:30 - 16:00 Lecture
16:30 - 18:00 Exercise Session
19:00 Joint Dinner


10:00 - 11:00 Lecture
11:00 - 12:00 Designing Posters in Groups
12:00 - 13:30 Lunch Break
14:00 - 15:00 Poster Session
15:30 - 17:30 Lecture & Group-building
17:30 Open Evening Program (including time for homework and self-organized dinner)


09:00 - 14:00 Time for free work, supervised by Inform
14:00 - 15:30 Presentations
15:30 - 16:30 FInal Discussion


Theory Speaker (Tuesday & Wednesday)

  Stefan Thurner Copyright: Stefan Thurner

Stefan Thurner, Medical University of Vienna

Stefan Thurner is full professor for Science of Complex Systems at the Medical University of Vienna, where he chairs Section for Science of Complex Systems. Since 2007 he is external professor at the Santa Fe Institute, since 2010 he is a part-time senior researcher at IIASA.

He obtained a PhD in theoretical physics from the Technical University of Vienna in 1995, and a PhD in economics from the University of Vienna in 2001. He held postdoc positions at Humboldt Universität zu Berlin and Boston University before he joined the faculta of the University of Vienna in 1999 and later Medical University. He obtaind his habilitation in theoretical physics in 2001. With his engagement with the Santa Fe Institute - he shifted his focus from theoretical physics to biological and complex systems, which are now his main scientific areas.

His work considers topics as broad as simulating cells in silico, measuring and modeling the dynamics of social human interactions, and analyzing and understanding network formation. In his lectures at the ORM Summer School, he will particularly consider the relationship of complexity and uncertainty in networked phenomena, such as financial markets.

Title: Systemic Risk and its Management

Systemic risk in financial markets arises either through synchronized behaviour of agents, or because of the interconnectedness of agents through financial contracts. We show that the systemic risk level of every agent in the system can be quantified by simple network measures. With actual central bank data for Austria and Mexico we are able to compute the expected systemic losses of an economy, a number that allows to estimate the cost of a crises. We can further show with real interbank data that is possible to compute the systemic risk contribution of every single financial transaction to the financial system. We suggest a simple financial transaction tax that taxes the systemic risk contribution of all transactions. This tax provides an incentive for market participants to trade financial assets in a way that effectively restructures financial networks so that contagion events become impossible. With an agent based model we can demonstrate that this Systemic Risk Tax practically eliminates the network-component of systemic risk in a system. We further discuss a series of policy relevant lessons learnt from the agent based model, that is one of the first to model the financial and the real economy in an inherently integrated way.

  Speaker Marc Uetz Copyright: Marc Uetz

Marc Uetz, University of Twente

Title: Stochastic Scheduling & Approximation Algorithms

Stochastic scheduling models are intriguing nonstandard combinatorial optimization problems, with partly surprising or counterintuitive properties.
In this set of two lectures we review the stochastic version of a few classical machine scheduling models, study the structure of solutions, some phenomena, and the difficulties in solving such problems. The main part of the lectures focusses on the design of approximation algorithms for the solution of such problems, mainly using linear programming techniques.
The design of approximation algorithms means to compute solutions, here scheduling policies, along with worst-case performance guarantees. The main difficulty lies in the fact that the benchmark to compare with, that is, optimal scheduling policies are highly complex and largely unknown. Hence the necessity to work with linear programming relaxations of stochastic scheduling problems. While proving worst-case performance guarantees is interesting in itself, it has the additional feature of providing also a lot of insight into the the problems and their solution. A set of exercises tailored along the topics of the lectures help to familiarise with the problems and techniques, and will (partly) be discussed in a tutorial session.


Praxis Speaker (Thursday & Friday)

  picture of Christopher Adler Copyright: Christopher Adler

Christopher Alder, Deutsche Lufthansa

Christopher Alder is senior manager in the Department of Revenue Management Strategy and Development at Deutsche Lufthansa since 2011. In his position, he conceptualizes new revenue management approaches to be implemented with Deutsche Lufthansa, interfaces with the relevant operational and strategic departments, and supervises doctoral students in research cooperations.

He received his PhD in mathematics from Clausthal University in 2010, after working in a research project on airline revenue management with Deutsche Lufthansa and Air Canada. His dissertation focused on revenue management with regard to dynamic pricing for parallel flights. Before re-joining Deutsche Lufthansa, he worked as a risk analyst at Hewitt-Aon.

His work considers topics in commercial strategy such as demand forecasting, price and availability optimization, multi-criteria revenue management and analyst training. In his seminar at the ORM Summer School, he will highlight current relevant topics in Lufthansa commercial IT as well as the role of uncertainty in these areas. During his practice seminar, students will have the opportunity to work on real-world case studies and present their work to a subject matter expert.

  picture of Ulrich Dorndorf Copyright: Ulrich Dorndorf

Ulrich Dorndorf, INFORM, Aachen

As Chief Technical Officer of INFORM, Ulrich Dorndorf has successfully applied advanced optimization methods in many different application areas, including production and inventory management.

During his seminar, students will investigate techniques for dealing with uncertainty in the area of inventory management. Starting from a classic deterministic optimisation model, the dynamic version of the economic lot sizing model, they will study different modelling and solution approaches from stochastic and robust optimization and relate them to industrial practice. Working in small groups, they will have opportunities for discussion and for presenting their results to an expert.