3 présentations

• 
  08h45 - 09h10

  Choice-Based Revenue Optimization and Intelligence Applied to the Railway Industry

  • Morad Hosseinalifam, prés., Expretio Technologies

  We present a new-generation of Revenue Optimization tools to meet the demands of railway operators. We integrate powerful customer behavior models into a large-scale optimization framework to face this dynamic and competitive environment. The framework developed by Expretio allows operators to optimally manage their seat inventory and set fares, as well as other product attributes, across channels and customer profiles while taking into account passenger purchase habits and competitors’ actions.

• 
  09h10 - 09h35

  Railway Demand Forecasting Using Machine Learning Approaches

  • Neda Etebarialamdari, prés.,
  • Gilles Savard, Polytechnique Montréal
  • Miguel F. Anjos, GERAD, Polytechnique Montréal

  In railway industries, demand forecasting is the estimation of the number of passengers aiming to travel by train with a specific itinerary. Railway uses this information to satisfy their demands and maximize the revenues. In this study, we present detailed analyses of applications of various machine learning algorithms and preprocessing techniques to predict the future bookings in railway industries in two different aggregation levels. As a result, stacked generalization method combined with proper preprocessing techniques outperformed other approaches at both levels. We successfully achieved 11% Mean Absolute Percentage Error for level-1 aggregation and 18% Weighted Absolute Percentage Error for level-2.

• 
  09h35 - 10h00

  Products closing times optimization for choice-based revenue management

  • Thibault Barbier, prés., Polytechnique Montréal
  • Gilles Savard, Polytechnique Montréal
  • Miguel F. Anjos, GERAD, Polytechnique Montréal
  • Fabien Cirinei, Expretio

  Choice behavioral aspect is now paramount to network revenue management. Tackling both "buy-up" and "buy-down" phenomena, choice-based models provide more accurate and robust solutions, thus generating better revenue. Most choice-based models optimize revenue by adjusting how much time each set of products must be offered over the booking process. This approach was introduced by the choice deterministic linear program (CDLP) and has been extensively developed since then to reduce resolution time and to tackle different choice behavior. We offer a new approach for choice-based network revenue management that finds when to stop selling each product during the booking process. Numerical experiments on literature instances demonstrate promising performance. Our model quickly returns a directly implementable control that often gives better revenue when simulated than traditional products sets approaches.