Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MB32

participation (supply), where interactions exist in that a service quality improvement of the integrated transit-ride-sourcing will attract more users for the combined mode and thus create incentives for more driver participation. An elastic demand stochastic user equilibrium model is developed and shows overall service improvement. 2 - Integrating Ride-shared Mobility-on-Demand (MOD) System with Public Transit Liu Xu, PhD Candidate, University of Maryland-College Park, 1173 Glenn L. Martin Hall,, College Park, MD, 20742, United States, Xinlei Zhang, Ali Haghani Seamless integration of ride-shared MOD service with public transit may offer sustainable and convenient multi-modal transportation. This paper proposes a partition-based real-time network optimization model for the hybrid system that capable of (i)scales to a large number of trips; (ii) incorporates dynamic ride- sharing to MoD services and (iii)generates optimal routes for real-time demands. To enhance applicability, the proposed model accounts for transit frequency, waiting time, MoD vehicle capacity and fleet size, travel delay and operations cost for vehicles. Extensive experimental results show a significant reduction of vehicle empty miles and a boost to mobility enhancement. 3 - Interoperable Smart Card Data Management in Public Mass Transit Stefan Voss, Professor, University of Hamburg, Institute of Information Systems (IWI), Von-Melle-Park 5, Hamburg, 20146, Germany, Filip Covic Due to a lack of shared practices of deployment, installation and application, smart ticketing is often built on proprietary specifications limiting their scope of integration and compatibility among each other. Therefore, interoperability in public mass transit has become a central aspect of e-ticketing. We develop a standardised process on how to handle smart card data in an interoperable environment. We present a unified approach where data mining tools and model applications can be tested and implemented in every region embedded in the integrated network. Our Interoperable Smart Card Data Chain provides a continuous procedure for standardised data handling and management. n MB32 North Bldg 222B Data-driven Freight Logistics Sponsored: TSL/Freight Transportation & Logistics Sponsored Session Chair: Satya Sarvani Malladi, Georgia Institute of Technology, Atlanta, GA, 30318, United States Co-Chair: Shahab Derhami, Georgia Tech, 755 Ferst Drive, NW, Atlanta, GA, 30332, United States 1 - Algorithms for Travel Speed Prediction using Big Data Provided by Home Delivery Company Maha Gmira,École Polytechnique de Montréal, Montreal, QC, Canada Maha Gmira, Canada Excellence Research Chair in Data Science for Real-Time Decision-Making, Montreal, QC, Canada, Michel Gendreau, Andrea Lodi, Jean-Yves Potvin In an urban environment, vehicle speed and consequently travel time can be highly variable, capturing time-varying features when modeling travel speeds provide an immediate benefit to commercial transportation companies that distribute goods. We present a travel speed prediction methodology based on data collected from mobile location devices installed inside commercial delivery vehicles. An analysis is conducted using unsupervised learning to cluster data, dimensionality reduction techniques, imputation methods and a Long Short Term Memory neural network to forecast travel speeds. 2 - Adoption of Electric Trucks in Freight Transportation Osman Alp, University of Calgary, Haskayne School of Business, 2500 University Drive, Scurfield Hall 120, Calgary, AB, T2N1N4, Canada, Tarkan Tan, Maximiliano Udenio We analyze a fleet management problem of a freight transportation company, which aims to adopt a fleet of sustainable vehicles in a finite planning horizon. In particular, we optimize the transformation of their fleet into a mixed fleet of electric and diesel trucks. Our model also optimizes for the choice of charging technology as well as the number of charging facilities and their capacities. The objective of the decision maker is to minimize the investment costs for vehicles and the charging infrastructure, and the carbon emission related costs. We present a numerical study to generate insights about how to facilitate and incentivize the adoption of sustainable vehicle technologies in practice.

n MB30 North Bldg 221C Joint Session TSL-FAC/SOLA: Strategic Location/Allocation of Resources Sponsored: TSL/Facility Logistics Sponsored Session Chair: Felipe Aros-Vera, Ohio University, Athens, OH, 45701, United States 1 - Optimal Districting of Disaster Areas Johanna Amaya Leal, Iowa State University, Ames, IA, United States, Jose Holguin-Veras, John E. Mitchell This paper describes a methodology to solve the problem of districting a disaster area among relief groups in post-disaster situations. Building upon formulations that minimize the total social costs of the operation, a districting strategy is proposed to expedite the flow of critical supplies to demand points. The model considers the resources the relief groups, capacity of sites, and the needs of the population to minimize the human impacts of the disaster 2 - The Integration of Research and Development, Manufacturing and Distribution in Humanitarian Supply Chains Nico Vandaele, KU Leuven, Naamsestraat 69, 0419.052.173, Leuven, 3000, Belgium, Catherine Jenny Decouttere, Kim De Boeck, Stef Lemmens, Mauro Bernuzzi Based on both our academic as well as our field work we elaborate on the importance of interconnecting R&D, manufacturing and distribution activities in vaccine supply chains. The ultimate goal of immunization and being prepared for outbreaks is highly dependent on a number of design and managerial decisions taken decades before the operations materialize and when responsiveness becomes key. We illustrate this with evidence from our modelling work in vaccine manufacturing as well as in local distribution settings in some sub-Saharan countries. 3 - A Framework to Assess Location of Micro Grids to Increase Energy Grid Resilience Felipe Aros-Vera, Ohio University, 277 Stocker Center, 1 Ohio University, Athens, OH, 45701, United States, Shayne Gillian, Austin Rehmar, Landon Rehmar This paper develops a framework to assess the location of microgrids to increase resilience of an electrical power system and mitigate the impact of catastrophic grid failure in disaster situations. Using optimization models and network vulnerability analysis, the framework determines where to locate a set number of microgrids to provide maximum benefit to a population. This paper uses a case study of the 2017 hurricane season that caused catastrophic grid failure on the island of Puerto Rico to evaluate the allocation of microgrids on critical infrastructure. 4 - A Stochastic Model to Allocate Water in Post-disaster Environments Diana Ramirez-Rios, Rensselaer Polytechnic Institute, 134 25th Street, Troy, NY, 12180, United States, Sofia Perez-Guzman, Trilce Encarnacion, John E. Mitchell, Jose Holguin-Veras This study develops a stochastic mixed integer programming formulation for the allocation of water when there exists an uncertainty of time the population will return to normalcy. In the case of water deprivation, this commodity is usually expected during the first days after the disaster occurs, and it is known that if a person lacks water, after 5 days they will die. This formulation considered five scenarios related to the days expected for return to normalcy. Results from a case study provide an idea of the magnitude of human suffering in distribution of aid during post-disaster response. n MB31 North Bldg 222A New Data, Models, and Technology for Public Transportation Sponsored: TSL/Intelligent Transportation Systems (ITS) Sponsored Session Chair: Stefan Voss, University of Hamburg, Institute of Information Systems (IWI), Von-Melle-Park 5, Hamburg, 20146, Germany 1 - A Stochastic User Equilibrium Model for Integrated Transit and Ride-Sourcing Services Yufeng Zhang, University of Minnesota, 500 Pillsbury Drive S.E., Room 175, Minneapolis, MN, 55455, United States, Alireza Khani Integrating transit and ride-sourcing services can improve mobility by providing solutions for the “first/last mile problem. This study explores the equilibrium between the demand for the integrated system and the ride-sourcing drivers’

159