Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WD52

4 - Data-driven Decision Making in Last Mile Delivery Problems Sami Serkan Zarik, PhD Candidate, Eindhoven University of Technology, School of Industrial Engineering (Pav. F08), P.O. Box 513, Eindhoven, 5600 MB, Netherlands, Lucas Petrus Veelenturf, Tom Van Woensel, Gilbert Laporte The recent increase in online orders leads to logistical challenges such as low hit rates (number of successful deliveries) and unattended deliveries. In this paper, we consider exact and heuristic approaches in order to solve last mile delivery problems in which customers’ attendance probability data is considered for each of the time buckets in the planning horizon. We aim to improve the hit rate by considering routing and scheduling decisions. We develop an efficient heuristic approach by iteratively solving routing and scheduling problems. Numerical results will be presented. 5 - Partnership Uncertainty and Jugaad Achal Bassamboo, Northwestern University, 2001 Sheridan Road, Evanston, IL, 60208, United States, Sanket Patil, Prateek Raj The paper models how partnership uncertainty influences the level of investment in new projects. When businesses can delay their decisions to invest and instead search for an outside option (do Jugaad) uncertainty in commitment can emerge. We find that Jugaad can lead to a tragedy of commons, where it is individually rational to expand the option set, but such a large option set reduces commitment to new projects and hampers overall level of investment. n WD52 North Bldg 231C Practice- Transportation Freight II Contributed Session Chair: Allan Larsen, Technical University of Denmark, DTU Management Engineering, Building 424, DTU, Lyngby, DK-2800, Denmark 1 - Truck Appointment Systems Considering Impact to Drayage Truck Tours Mohammad Torkjazi, University of South Carolina, Columbia, SC, 29205, United States, Nathan Huynh, Samaneh Shiri This study proposes a novel approach for designing a Truck Appointment System (TAS) intended to serve both the marine container terminal operator and drayage operators. The aim of the proposed TAS is to minimize the impact to both terminal and drayage operations. 2 - Design and Operation of China Railway Express under Market Competition Yingzi Peng, Tsinghua University, Beijing, 100084, China, Lefei Li As an important symbol of “The Belt and Road transportation interconnection, China Railway Express has obtained a great increasement in the number of commodity freight. Most mathematical models of intermodal network design have been developed as stakeholder of an intermodal operator. In this research, we develop a network design and operation model in order to help China Railway Express, as a carrier, to solve for better freight routes and schedules considering competition with other carriers. The model is formulated as a bilevel nonlinear integer program, which is difficulty to solve. 3 - Design of a Relay Network in Long-haul Transportation Amin Ziaeifar, Southern Methodist University, Dallas, TX, 75205, United States, Halit Uster We present a new model to strategically design a relay network for long-haul transportation by considering alternative routings. We devise a Benders decomposition based solution algorithm that is enhanced by strengthened benders cuts, heuristics, and surrogate constraints to solve it efficiently. We present the computational results to demonstrate the efficiency of our algorithm on large-scale instances. 4 - Freight Demand Modeling in Bangladesh: A World Bank Project This presentation explains the freight modeling performed as a part of a project funded by the World Bank group. Freight generation, freight trip generation were modelled at district level in Bangladesh, using the survey conducted by the team. Freight Origin-Destination Synthesis, a process of obtaining the loaded and empty flows between the districts was performed, using the confidential Census datasets, road network, link travel time estimated using the GPS data, and the traffic count data for year 2013. Imports and export traffic flows were modeled using the Customs data. A doubly constrained gravity model is used for trip distribution, and Noortman and van E’s model is used for empty trips. 5 - Using Electric Vehicles for Commercial Urban Transports Allan Larsen, Professor, Technical University of Denmark, DTU Management Engineering, Building 424, DTU, Lyngby, DK-2800, Abdelrahman Ismael, Rensselaer Polytechnic Institute, Troy, NY, 12180, United States, Jose Holguin-Veras, Lokesh Kumar Kalahasthi, Wilfredo Yushimito

Electric vehicles (EVs) are facing a rapid development enabling such vehicles to be used in commercial transport. The EU project, EUFAL (Electric Urban Freight and Logistics), sets out to examine the potentials of using EVs in city logistics. This presentation will provide an overview of the planning and management problems met when implementing EVs in urban freight transport. Two real-life optimization problems dealing with routing EVs for collection of blood samples from private physicians and the fleet management of electric service vehicles visiting construction sites will be introduced. n WD53 North Bldg 232A Behavioral Operations II Contributed Session Chair: Hsuanwei Chen, San Jose State University, One Washington Square, San Jose, CA, 95192, United States 1 - Decision Making under Uncertain Demand and Uncertain Supply Somak Paul, The Ohio State University, Columbus, OH, 43202, United States, Elliot Bendoly, Nathan C. Craig There has been little empirical effort to distinguish the implications of demand and supply uncertainty separately on inventory decisions when both are present. In this study we design a controlled laboratory experiment to investigate the impact of demand and supply variability, as well as feedback frequency and experience on the optimality of the inventory decisions made. While we find the impact of demand uncertainty to be in line with standard assumptions, the impact of supply uncertainty appears counter-intuitive. We find that supply variability has a dampening effect on order deviation from the optimal and imposes limits on Jaime Andr s Castañeda, Universidad del Rosario, CL 12 C. 6 25, School of Management, Bogotá, 111711, Colombia, John Vargas, Daniel Rodr guez, David Anzola, Nelson Gómez This study analyzes the bullwhip effect through an agent-based simulation approach. We simulate the Beer Game, where an agent simulates each echelon. The agents place their orders according to Sterman (1989)’s ordering rule. We systematically manipulate the behavioral parameters of the rule across the agents and use clustering techniques to determine what behavioral configurations are beneficial for the supply chain. 3 - How Market Power Impacts Investment in Supply Chain Theory and Experiment Jingjie Su, University of Texas at Arlington, Box 19437, Arlington, TX, 76019, United States, Kay-Yut Chen Consider a supply chain network with two players, OEM and CM. They have relationships both as downstream supplier and competitor in the market. Now, when OEM has an option to invest in CM, based on different investment cost and market power, the theory predicts the different decision. Behavior experiment differs. 4 - A Novel Approach to Propensity Score Matching Incohort Construction Forbehavioral Nudging in Health Care Ryan Croke, Data Scientist, NextHealth Technologies, Denver, CO, 80202, United States A novel approach to cohort construction in a health care application is presented. Propensity score matching is used as a baseline to construct a control group to a given trial set. The experiment is aimed at illiciting positive health decisions by insured consumers and with large and rich datasets. In an effort to ameliorate the concerns of various researchers a hybrid method is used that optimizes over the member attributes as well as the propensity score. 5 - A Semantic Analysis of Super Bowl Word of Mouth on Tourism Social Media Hsuanwei Chen, Assistant Professor, San Jose State University, One Washington Square, San Jose, CA, 95192, United States, Yinghua Huang This study investigates how tourism social media word-of-mouth changes before and after a mega event. Using Super Bowl 2016-2018 as an example, semantic analysis is applied to explore the popular topics among tourists’ interactions on Twitter. Specifically, two questions will be answered: 1) what do tourists discuss about Super Bowl and its host destination on Twitter? and 2) how do tourists’ topics of interest evolve as the event progresses? The findings will help the destination event planners to better understand tourists’ opinions and monitor the issues emerged from the trajectory of the event. the extent to which experience benefits optimal ordering choice. 2 - Ordering Behavior in the Beer Game: An Agent-based Simulation Approach

Denmark, Dario Pacino, Michael Bruhn Barfod, Jonas Mark Christensen, Satya Sarvani Malladi

512