Graduiertenkolleg: Methods for Discrete Structures

Deutsche Forschungsgemeinschaft
faculty | junior-faculty | postdocs | students | associate students | former students | former associate students
|
locations | Term schedule | history
|
predoc-courses | schools | block-courses | workshops
partners


Monday Lecture and Colloquium


Monday, June 25, 2012

Technische Universität Berlin
Institut für Mathematik
Str. des 17. Juni 136
10623 Berlin
room MA 041

Lecture - 14:15

Xavier Gandibleux and Anthony Przybylski -
University of Nantes

Multi-Objective Combinatorial Optimization

Abstract:
The development of efficient exact algorithms for Multi-Objective Combinatorial Optimization (MOCO) problems is a central research topic in operations research. Major advances have been recorded these last 10 years. One example is the two phase method which is now able to deal with situations presenting more than two objectives. Another example is the definition of the multiobjective branch and bound principle and its successful applications on several optimization problems. Nevertheless real-life applications complexify the picture. Large scale multiple objective combinatorial optimization problems, and/or formulation combining continuous and discrete variables have to be considered.
This talk introduces some recent advances and open questions in that field, and develops an example of recent exact algorithms proposed.




Colloquium - 16:00

Kai-Simon Goetzmann - Technische Universität Berlin


Reference Point Methods and Approximation in Multiobjective Optimization

Abstract:
The basic concept in multicriteria optimization is Pareto optimality: a solution is Pareto optimal if improving one objective is impossible without worsening another. In general, however, the number of Pareto optimal solutions is exponential. We study a concept called reference point methods. A reference point solution is the solution closest to a given reference point in the objective space. This method gives a single Pareto optimal solution balancing the criteria. We also establish an algorithmic link to approximate Pareto sets: finding approximate reference point solutions is equivalent to approximating the set of Pareto optimal solutions.



Letzte Aktualisierung: 11.06.2012