Computational Integer Programming (MA8034)
- Dates: March 30 - April 3, 2020 (Compact Course)
- Lecturer: Timo Berthold (FICO, Zuse Institute Berlin)
- Audience: Master students, PhD students, and postdocs interested in mathematical optimization or operations research, in particular integer programming
- ECTS Credits: 3 (with exam on April 3)
- Location: This course will be held as an online course.
- Contact: Stefan Weltge
This compact course provides a comprehensive understanding of computational methods for mixed-integer programming (MIP). The main goal is to obtain state-of-the-art knowledge that allows to develop better MIP models.
Upon successful completion of this compact course, you can describe and prove the correctness of the fundamental algorithms: simplex, branch-and-bound, cutting plane separation. You will understand which additional algorithmic enhancements are required to make those algorithms applicable for practically relevant problems. You will be able to argue which heuristic decisions a MIP solver can take and what structural features it typically exploits.
As a prerequisite, basic knowledge in combinatorial optimization is expected, in particular graphs (trees, paths, matchings, flows), basics on algorithms and complexity theory (running time, the classes P and NP), and linear & integer programming.
- Achterberg, Constraint Integer Programming (Part II: Mixed Integer Programming), 2007, http://dx.doi.org/10.14279/depositonce-1634
- Kallrath, Algebraic Modeling Systems (Chapter 5: Xpress-Mosel), 2012
About the Lecturer
Timo Berthold is a leading expert in the development of state-of-the-art approaches for solving general mixed integer nonlinear programs.
Timo is a Technical Editor of Mathematical Programming Computation, and received several awards including the Klaus Tschira Award for Achievements in Public Understanding of Science and the Dissertation Award of the German Operations Research Society (GOR).