Research made public.

Our research in the areas of operations research, mathematics and computer science regularly appears in renowned scientific journals. Of course we also participate in international conferences and workshops. Our newest results are available in advance as preprints.

306
Vorträge
56
Preprints
116
Publikationen

Integer Programming and Combinatorial Optimization

Buch-Cover: A. Schrijver, Theory of Linear and Integer Programming
Schrijver's Buch-Cover inspirierte unser Lehrstuhllogo

We love discreteness.

We model optimization problems with highly expressive variables that lead to strong relaxations. In particular, we develop algorithms that exploit the structure of integer programs to solve them more efficiently. This often results in column generation/branch-and-price (CGBP) or other decomposition strategies. Our chair belongs to the leading international research groups in this field. A flagship project is the generic solver GCG which automatically performs Dantzig-Wolfe decompositions and applies CGBP.

Machine Learning meets Optimization

Statistik aus unserem Paper Bergner et al. (2015)
Statistiken aus Bergner et al. (2015)

Data-Based Algorithm Improvements.

The intersection of machine learning and optimization is a highly topical field of research. In machine learning also appear discrete optimization problems that are still rarely modeled via integer programs and solved to optimality. Inversely, machine learning approaches can help to enhance the understanding of optimization algorithms and provide new intuition for theoretic aspects. Our research focuses primarily on the latter of these two areas.

Präskriptive Analyse

A Puzzle
Ohne Optimierung fehlt Dir was!

From Data to Optimal Decision-Making.

All our practice projects are simultaneously projects of mathematical research. This is primarily due to the fact that we look for challenges that are still without or with only insufficient solutions. Nearly always new decision models must be developed and algorithms must either partially be adapted to the special structure of the application or be drafted completely from scratch. The practical tasks often result in more general theoretic questions whose answers in turn provide new findings for practice.

Research Fields of our Associates

Macht was mit Wahlkreisen

Sebastian Goderbauer: does weird stuff with voting districts

Macht was mit Wahlkreisen

Michael Bastubbe: a decent decom-poser

Software

Implementing Our Research.

Part of our research is performing computational experiments. To evaluate our ideas and algorithms, they need to be implemented. We often make the developed software available to the public.

Data

Benchmarks, Instances, Generators

We publish the data and instances upon which our research results are based. This allows the scientific community to compare to our methods and to comprehend and reproduce our results. We maintain libraries of benchmark instances for selected applications.

Partner des Lehrstuhls