If you are a student in computer science, engineering, management science or a related area and are considering to do a master thesis in our area, feel free to contact us. We are in particular interested in a thesis on the topics

• Solution of multicommodity network flow problems.
• Algorithms for generalized networks.
• Solution of set-partioning problems both pure problems and problems with side constraints.
• Rook pivoting for nonsymmetric LU factorization.
• Solution of quadratic and conic  problems with integer constrained variables.
• Solution of large sparse linear systems.

Even though an mixed-integer optimization problem is solved with a branch & bound algorithm then a heuristic for find good feasible solutions is useful. An interesting project is generalize the heuristics from linear to conic case. In particular the efficient generalization of the feasibility pump heuristic would be interesting.

Normally a specialized version of  primal simplex algorithm is consider the best algorithm for network flow problems.  However, as the result of Mittleman documents the a generic dual simplex algorithm does very well on network flow problems. Therefore, an interesting project is

• Present a specialized version of dual simplex algorithm for network flow problems.
• Implement the ideas in a suitable programming language such as C or C++.