4 Presentations

• 
  10:30 AM - 10:55 AM

  Planification opérationnelle d'un système adaptatif à la demande (DAS) à une seule ligne via une flotte de véhicules autonomes à capacité limitée

  • SAMEH GRAINIA, presenter, Université de Montréal
  • Teodor Gabriel Crainic, Université du Québec à Montréal
  • Bernard Gendron, Université de Montréal, CIRRELT
  • Sanjay Dominik Jena, Université du Québec à Montréal
  • Fausto Errico, Polytechnique Montréal

  Le défi de fournir des services de transport public efficaces dans les zones à faible densité et pendant les périodes de faibles demandes de transport a été relevé par le système semi-flexible : Demand Adaptive System (DAS), qui intègre des services flexibles au sein d'un système traditionnel. En effet, une ligne DAS sert un ensemble d'arrêts obligatoires dans une plage horaire prédéfinie et un ensemble d'arrêts facultatifs activés à la demande. Afin de combiner la flexibilité et l'efficacité du DAS avec les avantages attendus des véhicules autonomes, nous nous concentrons, dans ce travail, sur le problème de planification opérationnelle d’un DAS à une seule ligne via une flotte de véhicules autonomes à capacité limitée, dans le cadre d'un système d'information sous forme de messagerie instantanée pour la gestion dynamique des demandes. Nous proposons trois stratégies de gestion des demandes et nous développons deux modèles de programmation mixte en nombres entiers, correspondants à un parcours de véhicules circulaire et un parcours aller-retour respectivement. Les contraintes de capacité sont exprimées de deux manières différentes dans chaque problème et les modèles sont adaptables à chaque stratégie de gestion des demandes. L'objectif est de déterminer l'ensemble de demandes le plus rentable à servir, ses horaires d’embarquement et de débarquement et ses itinéraires; et déterminer les chemins des véhicules les plus économiques et ses horaires. Le problème est résolu par Cplex. Les résultats d'expériences numériques sont donnés et analysés.
• 
  10:55 AM - 11:20 AM

  Rebalancing problem for Bike-sharing Systems Based on Inventory Interval and Target

  • Jiaqi Liang, presenter, Polytechnique Montréal
  • Maria Clara Martins Silva, Polytechnique Montréal
  • Sanjay Dominik Jena, Université du Québec à Montréal
  • Daniel Aloise, Polytechnique Montréal

  Bike-sharing systems are blooming in numerous cities, which provide an environmentally friendly and convenient city traffic mode. However, the external environmental factor, e.g., weather and time, and internal stochastic characteristics of customer trips lead to the imbalance of stations and further unsatisfied demand. The operators employ trucks to rebalance bikes among the stations in an unbalanced state. However, the existing models and algorithms are not always applicable to real-world BSS operator rebalancing themes with hundreds of stations.

  In this paper, we propose two models, DROBDI and DROBDT, that combine a MIP model with inventory intervals, a mechanism widely used by BSS operators to facilitate rebalancing decisions. We performed experiments for synthetic trip data on different models in two planning frameworks, multi-periods and rolling. Then, we extend our proposal by applying a subset-selection algorithm to improve the solvability of optimization models with a large-scale network. Extensive numerical experiments are carried out on real-world data, provided by BIXI.
• 
  11:20 AM - 11:45 AM

  Towards Effective Rebalancing of Bike-Sharing Systems with Regular and Electric Bikes

  • Maria Clara Martins Silva, presenter, Polytechnique Montréal
  • Daniel Aloise, Universidade Federal do Rio Grande do Norte
  • Sanjay Dominik Jena, Université du Québec à Montréal

  The emerging demand for electric bicycles in recent years has led several bike-sharing systems around the world to adapt their service to a new wave of commuters. Many of these systems have included electric bikes in their network, without abandoning the use of regular mechanical ones. Nonetheless, the presence of two types of bike-sharing docks at the system's stations may impact how operations should be conducted to rebalance a BSS. Indeed, regular and electric bikes may present very different demands throughout the day, making harder the planning of such operations. In this paper, we propose a model to provide tailored rebalancing recommendations for BSS with shared regular and electric bikes. According to our results, our model managed to reduce the lost demand by approximately 81\% with respect to the rebalancing strategy used by the real-world studied BSS, and by approximately 66\% in comparison with a state-of-the-art model, while performing a comparable amount of rebalancing operations in the stations of the system.
• 
  11:45 AM - 12:10 PM

  Optimizing costs and carbon emissions for courier deliveries in Montreal through electric vehicle use and warehouse sharing

  • François Sarrazin, presenter, Université de Montréal
  • Martin Trépanier, Polytechnique Montréal, CIRRELT
  • Matthieu Gruson, UQÀM - CIRRELT

  City logistics has become more important over the years as a greater share of the world’s population lives in urban areas. This is especially true in an area of increasing use of e-commerce and greater awareness of the need to reduce transportation externalities to combat global warming. In this paper, we evaluate the potential for reducing costs and CO2 emissions in the courier delivery industry by sharing warehouses between different deliverycompanies. We also consider a greater use of cargo bikes (CBs),which are more mobile and eco-friendlier. We test a cost reduction optimization model on data for the delivery of courier in the metropolitan area of Montreal, Canada. Results show a cost and GHG emissions reduction (of over 30%) when warehouses are shared between different courier companies and when CBs are used. We define a ratio between the maximum distance between a pickup or delivery site and the depot (warehouse) used and the total distance travelled. Results show that such a ratio has a significant impact on the gains brought using CBs. The maximum speed of vehicles also has a significant impact on the results.