Journées de l'optimisation 2014

                             Incluant une Journée industrielle de l'optimisation

                                              HEC Montréal, 5 - 7 mai 2014


HEC Montréal, 5 — 7 mai 2014

Horaire Auteurs Mon horaire

WA5 Transport et gestion du bois / Transportation and Wood Handling

7 mai 2014 09h00 – 10h40

Salle: Nancy et Michel-Gaucher

Présidée par Bernard Gendron

4 présentations

  • 09h00 - 09h25

    Models and Mathematical Decomposition for Large-Scale Location Problems in the Forestry Sector

    • Sanjay Dominik Jena, prés., Université du Québec à Montréal
    • Bernard Gendron, Université de Montréal, CIRRELT
    • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT

    We present mathematical models and algorithms for a complex location problem which is found in the forestry sector. Based on logging demands for the next five years, this problem investigates the optimal number, locations and sizes for new camps in order to host workers involved within logging activities and balance transportation, camp construction and camp relocation costs. The problem includes a very detailed cost structure. We compare different models and present heuristics based on two different Lagrangian decomposition approaches, capable to solve even very large instances. Case studies are provided based on data from a Canadian logging company. Given the generality of the problem, the presented models and algorithms may also be applied to other location problems typically found in the forestry sector.

  • 09h25 - 09h50

    Dock and Driver Scheduling in a Timber Transport Supply Chain

    • Gregory Rix, prés., Polytechnique Montréal
    • Louis-Martin Rousseau, Polytechnique Montréal
    • Gilles Pesant, Polytechnique Montréal

    We present an inventory-routing problem with pickups and deliveries that arises in the forest products industry. A heterogeneous vehicle fleet must be scheduled to meet driver constraints, and also synchronized with loading equipment at supply and demand points. A branch-and-price based heuristic is developed to to generate near-optimal solutions under a practical time limit.

  • 09h50 - 10h15

    A Heuristic Approach for a Tactical Harvest Planning Problem under Market Price Uncertainty

    • Géraldine Gemieux, prés., Université de Montréal
    • Bernard Gendron, Université de Montréal, CIRRELT
    • Jacques Ferland, Université de Montreal
    • Andres Weintraub, University of Chile

    We consider a tactical harvest planning problem driven by final products demands where the objective is to maximize the net present value. However the prices of final products are under uncertainty. This uncertainty is modeled by a scenario tree. Each realization of scenario has an associated probability. The problem is then to find solution for each scenario that maximizes the expected value of the NPV of each scenario and respect the non anticipativity principle. This principle ensures that decision are only taken considering parameters available at the moment. Those constraints increase the difficulty to solve the formulated stochastic mixed integer programming problem. We use a heuristic solution process based on a decomposition method by scenarios (progressive hedging). Preliminary results in Chilean case study will be presented.

  • 10h15 - 10h40

    Real-Time Transportation and Logistics Systems Planning and Control

    • Amine Amrouss, prés., Université de Montréal
    • Michel Gendreau, Polytechnique Montréal
    • Bernard Gendron, Université de Montréal, CIRRELT

    Once a transportation plan is derived for one day at the operational planning level and the wood transportation from forest areas to mills starts, several unforeseen events may occur, events that perturb planned trips (e.g., because of weather conditions, forest fires, or the occurrence of new loads). When such events take place and the demand at mills can no more be satisfied if the daily plan is still followed, the latter must be reoptimized taking into account the new changes. The time available for computing a good choice is limited. Hence, there is a need to develop systems that are capable of supporting the dispatcher in real-time.
    We present the first steps towards developing a tool that will allow forest companies to quickly redraw their truck schedule to respond to any event that has perturbed the original operational truck schedule.