**November 15, 8:30-09:00 – Registration
November 15, 9:00-17:45 – Scientific Program
November 15, 18:15- Drink-Buffet
**

**Scientific Program (amphitheater located in building 34):
**Last update: November 18, 2018

**09:00 – 09:10: Opening: **Xavier Gandibleux, Claude Jard

**09:10 – 10:00: Keynote I
**Chair: Stefan Ruzika (Technischen Universität Kaiserslautern, Germany)

Vincent T’Kindt (Université de Tours, France). **Quantifying the hardness of the enumeration of Pareto optima: a theoretical framework with application to scheduling problems.** (slides)

(Abstract)

In this talk we will focus on recent theoretical developments that intend to establish some quantification, in the worst-case, of the complexity of enumerating the set of Pareto optima when this cannot be done in polynomial time. This follows early works, in the field of Computer Science, on the proposal of “Exponential Time Algorithms” (ETA) for single criterion NP-hard optimization problems. Such an algorithm has calculable worst-case time and space complexities which are as low as possible (and always better than brute-force algorithms). Then, such complexities give an hint on the tractability of an NP-hard problem. For example, it is now known that the maximum independent set problem in graphs can be solved in O*(1.1996^n) time in the worst-case with n the number of vertices.

Up to now, no work has addressed this issue for multiobjective optimization problems. We will remedy to that by introducing the basic notions related to ETA and show the existence of such algorithms for some multiobjective hard scheduling problems.

**10:00 – 10:10 Break**

**10:10 – 11:25 Session 1: Algorithms**

Chair: Kathrin Klamroth (Bergische Universität Wuppertal, Germany)

10:10 – 10:35: Sune Lauth Gadegaard (Aarhus Universitet, Denmark). **Finding all efficient solution to a bi-objective combinatorial optimization problem.** (slides)(Abstract)

10:35 – 11:00: Stefan Ruzika (Technischen Universität Kaiserslautern, Germany). **Can we approximate a weight set decomposition?** (slides) Joint work with Pascal Halffmann (Technischen Universität Kaiserslautern).

(Abstract)

In this talk, we introduce the idea of an approximation for computing the weight set decomposition. To the best of our knowledge, this has not been done before, hence we lay the foundation for further development in this research.

First, we state what we can understand as an approximate weight set decomposition and we list requirements and conditions from which we construct a proper definition. Further, we present various measures in order to state the quality of the approximation such as an approximation factor. Last, we present various approaches of an approximation algorithm that decomposes the weight set. We examine this algorithm regarding correctness, approximation quality, convergence rate, running time and bounds and give some illustrative examples.

11:00 – 11:25: Luís Paquete (University of Coimbra, Portugal). **Some results for the Hypervolume Subset Selection Problem. **Joint work with Miguel M. Duarte (University of Coimbra), José R. Figueira (University of Lisbon), Carlos M. Fonseca (University of Coimbra), Ricardo J. Gomes (University of Coimbra), Andreia P. Guerreiro (University of Coimbra), Tobias Kuhn (University of Kaiserslautern), Stefan Ruzika (University of Kaiserslautern).

(Abstract)

**11:25 – 11:50 Coffee Break**

**11:50 – 12:40 Session 2: Representations**

Chair: Matthias Ehrgott (Lancaster University, UK)

11:50 – 12:15: Martin Philip Kidd (Technical University of Denmark, Denmark). **Equidistant representations: connecting coverage and uniformity in biobjective optimization.** (slides) Joint work with Richard Martin Lusby and Jesper Larsen (Technical University of Denmark).

(Abstract)

12:15 – 12:40: Kerstin Daechert (Fraunhofer Institute for Industrial Mathematics ITWM, Kaiserslautern, Germany). **Obtaining representations for continuous optimization problems.** (slides)

(Abstract)

**12:40 – 14:00 Lunch Break**

**14:00 – 15:15 Session 3: Applications I**

Chair: Lars R. Nielsen (University of Aarhus, Denmark)

14:00 – 14:25: Antoine Kerbérénès (Université Paris-Dauphine and Naval Group Research, France). **Multi-Objective Combinatorial Optimization for Weakly Coupled Systems.** Joint work with Daniel Vanderpooten (Université Paris Dauphine), and Jean-Michel Vanpeperstraete (Naval Group Reasearch).

(Abstract)

14:25 – 14:50: Onur Tanil Doganay (Bergische Universität Wuppertal, Germany). **Gradient-Based Biobjective Shape Optimization of Ceramic Components: Probability of Failure versus Cost.** (slides) Joint work with Johanna Schultes, Camilla Hahn, Hanno Gottschalk, Kathrin Klamroth, Michael Stiglmayr (University of Wuppertal).

(Abstract)

This changes, if deterministic reliability criteria are replaced by a probabilistic criterion, namely the probability of failure (PoF), which has been introduced recently to the field of ceramic design and low cycle fatigue.

Here we present a finite element based first discretize, then adjoin approach for the calculation of shape gradients (sensitivities) with regard to the PoF of ceramic designs. This is applied to the simultaneous minimization of the PoF and the volume of a 2D ceramic rod in a biobjective shape optimization problem.

We have implemented a biobjective descent algorithm and compare it with a classical weighted sum approach.

14:50 – 15:15: Nicolas Dupin (Université de Lille, France). **Scheduling maintenances of nuclear power plants, from 2-stage robust programming to multi-objective optimization** (slides). Joint work with El-Ghazali Talbi (Université de Lille).

(Abstract)

**15:15 – 15:30 Break**

**15:30 – 16:20 Session 4: Applications II**

Chair: Daniel Vanderpooten (Université Paris-Dauphine, France)

15:30 – 15:55: Nicolas Forget (Université de Nantes, France). **On two speeding-up techniques for the computation of multi-objective shortest paths with a label setting algorithm**. (slides) Joint work with Xavier Gandibleux, Didier Robbes (Université de Nantes) and Matthias Ehrgott (Lancaster University).

(Abstract)

15:55 – 16:20: Clemens Thielen (Technischen Universität Kaiserslautern, Germany). **Duty Rostering for Physicians at a Department of Orthopedics and Trauma Surgery: Decision Support using Mathematical Optimization**. (slides)

(Abstract)

We present an integer programming model for the generation of duty rosters for physicians at a large department of orthopedics and trauma surgery that is used in practice since January 2016. Using real-world data from the practice partner, we evaluate the quality of the generated duty rosters with respect to several objective functions and additionally present ways of dealing with unpredictable disruptions (caused, e.g., by absence of physicians due to illness).

**16:20 – 16:45 Coffee Break**

**16:45 – 17:35: Keynote II**

Chair: Sophie Parragh (Johannes Kepler University-Linz, Austria)

Natashia Boland (Stewart School of Industrial & Systems Engineering, Georgia Institute of Technology, USA). **Criterion Space Search Algorithms for Solving Multiobjective Mixed Integer Linear Programs**.

(Abstract)

**17:35 – 17:45: Closing:** Xavier Gandibleux, Stefan Ruzika