Optimization Days 2019

HEC Montréal, May 13-15, 2019


HEC Montréal, 13 — 15 May 2019

Schedule Authors My Schedule

MB6 OR/MS Scientific Writing Activity - Methods and Applications in Logistics

May 13, 2019 10:30 AM – 12:10 PM

Location: Marie-Husny

Chaired by Marilène Cherkesly

4 Presentations

  • 10:30 AM - 10:55 AM

    The multi-objective multi-period location routing problem: A mobile clinic application

    • Rosemarie Santa González, presenter, UQÀM
    • Marilène Cherkesly, GERAD - Polytechnique Montréal
    • Marie-Eve Rancourt, HEC Montréal
    • Teodor Gabriel Crainic, CIRRELT - École des sciences de la gestion - UQÀM
    • Teodor Gabriel Crainic, CIRRELT - École des sciences de la gestion - UQÀM
    • Marie-Eve Rancourt, HEC Montréal
    • Marilène Cherkesly, GERAD - Polytechnique Montréal

    In this study we present the Multi Period Location Routing Problem (MLRP) formulation and illustrate it with a mobile clinic application. Mobile clinics give healthcare practitioners the ability to provide medical attention to populations that have limited access to healthcare. When conducting mobile clinic operations practitioners have to locate the depots, select the locations at which services will be offer, and design the schedule for the visits. There is more than one objective when planning a mobile clinic operation, therefore the problem is multi objective in nature. Due to the fact that medical treatments depend on the order and frequency in which they are administered this renders the problem as a multi period problem. The model proposed is tested on real life instances generated with the input of the Premier Urgence Internationale (PUI).

    Location Routing Problem, Healthcare, Humanitarian Logistics, Multiperiod, Multicriteria

  • 10:55 AM - 11:20 AM

    Logic-based Benders reformulations for integrated process configuration and production planning problems.

    • Karim Perez Martinez, presenter, HEC Montreal
    • Jans Raf, HEC Montréal
    • Yossiri Adulyasak, HEC Montréal

    This research addresses production planning problems where products of different types can be
    produced simultaneously according to a specific process configuration or pattern. The problem
    consists of determining the configurations to be used and the production level of each configuration
    to fulfill the demand at the minimum total cost, which typically includes setup costs, inventory
    holding or overproduction costs. We propose logic-based Benders reformulations and a branch-and-check algorithm to optimally solve this problem in different industrial contexts. The proposed methods outperform the
    benchmark approaches in the tested problems.

  • 11:20 AM - 11:45 AM

    An experimental study on learning convex parametric optimization programs via inverse optimization and machine learning

    • Elaheh Iraj, presenter,
    • Terekhov Daria, Concordia University

    We study the problem of learning from data that is generated by a parametric
    optimization process using inverse optimization. We reinterpret the
    applicability of inverse optimization for the purpose of learning and
    experimentally compare its predictive performance with machine learning
    algorithms: Random Forest, Support Vector Regression and Gaussian Process.
    Key Words: Inverse Optimization, Parametric Optimization, Machine Learning

  • 11:45 AM - 12:10 PM

    Routing hub location problem

    • Luiza Bernardes Real, presenter, IFMG
    • Ivan Contreras, Concordia University
    • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT
    • Ricardo Saraiva de Camargo, Federal University of Minas Gerais
    • Gilberto de Miranda, Federal University of Espírito Santo

    We introduce a variant for the routing hub location problem. We assume that each route may contain a mix of non-hub and hub nodes, commodity transfers can only be done at hubs and transportation costs are flow-dependent. A mathematical formulation is proposed and computational experiments are presented.

    Keywords: hub-location; vehicle-routing; network-design