The next D3 seminar will take place on the 8th of March at 4.00 pm in room A008.
Dr. Hatice Calik will présent her works on sharing optimization of vehicles with stochastic demands.
She candidates on the job of lecturer at Polytech (ex-ESSTIN).
Title: A Benders decomposition method for locating stations in a one-way electric car sharing system under demand uncertainty
Abstract: The increasing number of privately owned cars pushes the city managers to find solutions to the increasing pollution, traffic congestion, and parking problems in urban areas. Electric car sharing systems, which are based on shared use of vehicles owned by a company or an organization, have a high potential of reducing or eliminating these problems by limiting private car usage and ownership. Nevertheless, they need to be designed and operated in a way to provide high levels of accessibility and flexibility to be effective. In this study, we focus on the design of a one-way car sharing system with a fleet of identical electric cars. The system under consideration provides flexibility in the sense that the users are allowed to leave the car to a station different from the pick-up one and no pre-booking is enforced. This freedom of usage leads to an uncertainty in the demand. We approach the system from a strategic point of view and aim to decide on the location of stations and the initial number of cars available at each station in a way to maximize the overall expected profit. We introduce multiple demand scenarios to represent the demand uncertainty and formulate the problem as a mixed integer stochastic programming model. The profit function takes into account the expected revenue obtained from the served user requests and the fixed costs of opening stations and purchasing cars. In order to solve large-scale problems optimally, we further develop a Benders decomposition method based on our formulation. To improve the convergence speed of our algorithm, we strengthen our formulation with valid inequalities and introduce a stabilization procedure. We conduct computational experiments on problem instances obtained from real data based on Manhattan taxi trips. We are able to solve problems with 100 to 500 scenarios, each scenario including 1000 to 5000 individual customer requests, under high and low cost values and 5 to 15 minutes of accessibility restrictions, which is measured as the maximum walking time to the operating stations.
Acknowledgement: This is a joint work with Bernard Fortz from Université Libre de Bruxelles, Department of Computer Science, Brussels, Belgium. This research is conducted under e4-share (Models for Ecological, Economical, Efficient, Electric Car-Sharing) project funded by FFG, INNOVIRIS and MIUR via JPI Urban Europe.
Biography: Dr. Hatice Çalık is a post-doctoral researcher in the OPTIMIST group of LORIA since October 2017. She got her PhD from the Department of Industrial Engineering of Bilkent University in Ankara, Turkey in 2013. She worked as a post-doctoral researcher in CIRRELT and HEC Montréal in Canada in 2014 and in the Department of Computer Science of Université Libre de Bruxelles in Brussels, Belgium from 2015 to October 2017. Her research interests are on modelling and development of exact and heuristic methods for discrete optimization problems from location, transportation, and logistics applications.