Journées de l'optimisation 2018

HEC Montréal, Québec, Canada, 7 — 9 mai 2018

Horaire Auteurs Mon horaire

WA9 On the integration of machine learning and mathematical optimization II

9 mai 2018 10h30 – 12h10

Salle: Quebecor (80)

Présidée par Jean-Yves Potvin

4 présentations

  • 10h30 - 10h55

    Accelerating the optimization of aircrew rotations with machine learning

    • Yassine Yaakoubi, prés., Gerad
    • Francois Soumis, GERAD et Polytechnique
    • Simon Lacoste-Julien, MILA, Université de Montréal

    The optimization of crew rotations is a critical problem in air transport. So far, this problem has been handled by GENCOL: a solver that uses column generation to minimize costs while respecting collective agreements. We combine artificial intelligence and operational research to accelerate this optimization to achieve a better solution.

  • 10h55 - 11h20

    OR/ML for recommendation systems, retail assortments, and financial markets

    • Sanjay Dominik Jena, prés., Université du Québec à Montréal
    • Laurent Charlin, HEC Montréal
    • Gunnar Grass, HEC Montréal
    • Tolga Cenesizoglu, HEC Montréal

    This presentation focuses on two independent projects. The first project proposes to explore the synergies between recommendation systems and assortment optimization. While both applications have similar objectives, they have been treated mainly by distinct research and practitioner communities. We propose to exploit the synergies on both the application and the methodological level. The second project aims at finding optimal real-time order strategies for large institutional orders in financial markets, based on historical data-sets that account for multiple terabytes.

  • 11h20 - 11h45

    A network design problem arising in the restoration of the water supply system in Nepal

    • Selene Silvestri, prés., HEC Montréal
    • Gilbert Laporte, HEC Montréal
    • Marie-Eve Rancourt, HEC Montréal

    We propose an optimization approach for the community water network rehabilitation problem targeting remote populations affected by the 2015 Nepal earthquake. To this end, we describe the problem and a matheuristic developed to solve it. Data provided by the Red Cross and satellite imagery are used to generate solutions.  

  • 11h45 - 12h10

    Nouvelles approches pour la modélisation et la résolution de problèmes de livraisons à domicile

    • Jean-Yves Potvin, prés., Université de Montréal
    • Maha Gmira, Polytechnique Montréal
    • Michel Gendreau, Polytechnique Montréal
    • Andrea Lodi, Polytechnique Montreal
    • Thibaut Vidal, Pontifícia Universidade Católica do Rio de Janeiro

    Ce projet s’intéresse à la génération de tournées de véhicules pour des problèmes de livraisons à domicile. Une partie des travaux portera sur la modélisation de ces problèmes, particulièrement les objectifs poursuivis par les expéditeurs. L’autre partie s’intéressera au développement de méta-heuristiques permettant de produire des tournées de bonne qualité.