ὅδε οἶκος, ὦ ἑταῖρε, μνημεῖον ἐστιν ζῴων τῶν σοφῶν ἀνδρῶν, καὶ τῶν ἔργων αὐτῶν

THE DEVELOPMENT OF HYBRID HEURISTICS FOR COMBINATORIAL OPTIMIZATION PROBLEMS



Project approved by the Serbian Ministry of education, science and technological development and the Centre national de la recherche scientifique from France as a part of bilateral cooperation between two countries.

Dates: February 2016 - December 2017

Project number: 451-03-39/2016/09/09

Institutions:

from France: ESSEC Business School of Paris
from Serbia: Mathematical Institute SANU (Belgrade)

Project coordinators:

1. Ivana Ljubić, ESSEC Business School of Paris, Av. Bernard HirschB.P. 5010595021 Cergy Pontoise Cedex France (email: ljubic@essec.edu)
2. Tatjana Davidović, Mathematical Institute SANU, Knez Mihailova 36, 11000 Belgrade, Serbia (email: tanjad@mi.sanu.ac.rs)

Other project members:

from France

1. Fabio Furini, Laboratoire d'Analyse et Modélisation de Systèmes pour l'Aide à la DEcision (LAMSADE), Université Paris Dauphine, UMR-CNRS 7243
2. Laurent Alfandari, ESSEC Business School of Paris
3. Sébastien Martin, Laboratoire de Conception, Optimisation et Modélisation des Systèmes (LCOMS), Université de Lorraine

from Serbia

1. Tatjana Jakšić Kruger, Mathematical Institute SANU, Belgrade
2. Zorica Stanimirović, Faculty of Mathematics, University of Belgrade
3. Stefan Mišković, Faculty of Mathematics, University of Belgrade
4. Vladimir Ilin, Faculty of Technical Sciences, University of Novi Sad
5. Vladislav Maraš, Faculty of Transport and Traffic Engineering, University of Belgrade
6. Jelena Jovanović, Faculty of Computer Science, John Naisbitt University, Belgrade

Project goals: To establish cooperation between the two teams on the development of hybrid methods for solving difficult combinatorial optimization problems. These problems often occur in real life - in transportation, telecommunication, industry, economy, military, medicine, emergency systems, and many other areas. Usually, these problems are very complex, i.e., NP-hard. This means that, for the large-scale input data, the exact methods (which give exact optimal solution) cannot be directly applied since they are usually time and/or memory demanding. On the other hand, heuristic methods, i.e., approximate methods can not provide optimal solutions, sometimes not even the assessment of the quality for generated solutions, but they provide (suboptimal) solutions in short running times. Therefore, in the recent literature, hybrids of the two approaches are considered. There are two directions in the development of hybrid heuristics. The first is to start from a given real life problem and develop the most appropriate method that exploits the knowledge about the nature of the problem. The second direction aims to develop generalized methods that can be applied to various optimization problems. The aim of this research is to try to cover both directions, whereby, as the first case, we will consider the vehicle routing problems in river and road transport.