Keynote Speakers

Stefan Voß
Institute of Information Systems, University of Hamburg (Germany)
download extended abstract

Abstract: Popular matheuristics include the corridor method, POPMUSIC and generalized local branching. Common features use local search based candidate solutions by means of a fully-fledged optimization method generating optimal solutions over appropriately defined neighborhoods. That is, neighborhoods are constructed to be suitable domains for the optimization method used. We clarify differences and commonalities of these methods and exemplify by network design and maritime shipping.

Martin Josef Geiger
University of the Federal Armed Forces Hamburg (Germany)
download extended abstract

Abstract: We are fond of optimization/implementation challenges. First of all for one reason: they provide us researchers with a testbed for applying our methods, skills, and ideas, and come with impartial verification of the results by a third party, i.e. the challenge organizers. Moreover, the development of the solution approaches must be completed within a finite, typically short time horizon, and important factor when judging on the ‘practical applicability’ of a solution method. Even better, the problems are not chosen by the participants themselves. When putting everything together, some meaningful comparison should arise.
We may suspect that metaheuristics can successfully be applied in such scenarios. While this is, from our experience, true, their actual application is not so easy. This paper therefore collects some experiences with actual implementations for different optimization challenges that you may find helpful.