4 présentations

• 
  10h30 - 10h55

  Optimization of Train Timetables Adapted to Dynamic Demand Based on Riemann’s Sums

  • Eva Barrena Algara, prés., HEC Montréal
  • David Canca, Universidad de Sevilla
  • Leandro C. Coelho, Université Laval
  • Gilbert Laporte, HEC Montréal

  We study the non-cyclic railway timetabling problem adapted to dynamic demand, thus focusing on passenger welfare. We analyze its similarities with Riemann’s Sums and make use of the results of the latter in order to propose a metaheuristic which solves large instances of the problem within short computation times.

• 
  10h55 - 11h20

  Improving Branching in the Crew Pairing Problem with Credit Constrains

  • Frédéric Quesnel, prés., Polytechnique Montréal
  • Francois Soumis, GERAD et Polytechnique
  • Guy Desaulniers, GERAD - Polytechnique Montréal

  The airline crew scheduling problem is studied by many researchers. Usually, the problem is divided in two steps : the crew pairing and the crew rostering problems. This paper focuses on the former.

  We first show data relative to the airline crew scheduling problem, which will be published on the Internet. These are designed to be used as standard instances by other researchers, both on the crew pairing and the crew scheduling problem. We then present an original version of the pairing problem that includes constraints limiting the credit allowed for each base during the whole period.
  In some cases, these constraints increase difficulty to reach integer solutions. We present a systematic analysis of the behaviour of the branching phase according to the tighness of the constrains.

• 
  11h20 - 11h45

  A Heuristic Algorithm for Personalized Crew Scheduling Problem

  • Atoosa Kasirzadeh, prés., Polytechnique Montréal
  • Mohammed Saddoune, Polytechnique Montréal
  • Francois Soumis, GERAD et Polytechnique

  We present a set-covering formulation and an iterative heuristic algorithm for personalized integrated cockpit pairing and assignment problems. The objective is having as much as possible similar pairings between pilots and co-pilots to increase the schedules robustness, even if pilots and co-pilots schedules are different to satisfy their preferences. We use a solution approach based on column generation for this problem. The computational results are provided based on a major US carrier data set.

• 
  11h45 - 12h10

  Interactions Between Operations and Planning in Air Traffic Control

  • Thibault Lehouillier, prés., Polytechnique Montréal
  • Jérémy Omer, Polytechnique Montréal
  • Cyril Allignol, École Nationale de l'Aviation Civile
  • Francois Soumis, GERAD et Polytechnique

  Air traffic in Europe is predicted to largely increase over the next decades. In such a context, we present a study of the interactions between costs due to ground holding regulation and costs due to en-route air traffic control. With that in mind, a traffic simulator including the computations of regulation delays, aircraft trajectories and air conflict resolution is described. Through intensive simulations based on traffic forecasts extrapolated from 2012 historical French traffic data,regulation delays and avoidance maneuvers are computed assuming the current regulation or no regulation at all. The resulting costs analysis highlights the exponential growth of regulation costs that should be expected if the airspace capacity and the involved procedure do not change. Compared to this, the costs of air traffic control remain negligible whether regulation is performed or not. The analysis of controllers’ workloads however emphasizes the future need to combine automated tools assisting controllers with a regulation better adapted to bigger traffic volumes.