Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WB06

n WB06 North Bldg 122C Interface of Consumer Behavior and Retail Operations Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Stanley Lim, University of Cambridge, Cambridge, United Kingdom Co-Chair: Annibal Sodero, University of Arkansas, Fayetteville, AR, United States 1 - Operational Execution and POP Display Effectiveness: Evidence from Adoption of an IoT Technology Ioannis Stamatopoulos, University of Texas at Austin, McCombs School of Business, 2110 Speedway, B6000 Austin, TX 7870, Austin, TX, 78705, United States, Ashish Agarwal, Jacob Zeng POP displays are structures containing inventory outside of regular shelving. They are meant to be visually stimulating, are placed in prominent positions in retail stores, and their purpose is to excite spontaneous purchases. POP displays are perceived by suppliers to be highly effective and they invest in them heavily. However, poor operational execution is reducing their otherwise high effectiveness. We are interested in quantifying the extent of the problem, as well as exploring whether an IoT technology can improve operational execution. 2 - Using Transactions Data to Improve Consumer Returns Forecasting Guangzhi Shang, PhD, Florida State University, Erin Cassandra McKie, Mark Ferguson, Michael Galbreth Although an accurate returns forecast is the preliminary input into many decision support tools for managing returns, the development methods in this area received relatively little attention as compared with sales forecasting. We propose a new approach and benchmark its performance against a number of existing methods using two real world datasets. 3 - Free Shipping is Not Free: A Data-driven Model to Design Free-shipping Threshold Policies Joseph Xu, PhD, Carnegie Mellon University, Gerard P. Cachon, Santiago Gallino We provide a data-driven analytical model to (i) assess the profitability of an online retailer’s free shipping threshold policy, and (ii) offer recommendations to determine a suitable level of free shipping threshold policy to maximize profit. The model accounts for various aspects of customer behavior, such as strategically adding items to shopping basket to receive free shipping and changing the quantity of item returned. We calibrate our model to data from an online apparel retailer and find that its decision to offer a lower free shipping threshold reduce its profitability considerably. 4 - An Analysis of Shopping Behaviour at Warehouse-club Stores and its Store-network-density Implications Stanley Lim, University of Cambridge, Tempe, United Kingdom in the retail sector, we are aware of no empirical studies examining how significant this source of competitive advantage is for them. We use a quasi- natural experiment with households’ subscriptions to Costco stores as a treatment mechanism to contrast members’ purchasing behaviors at Costco stores against changes in behaviors at non-WC stores. We show that members’ weekly mileage accumulated per visit, dollar spent per visit, and dollar spent per mile to Costco stores exceeds by 7.4%, 29% and 20%, respectively, relative to non-WC stores. 5 - Inventory Allocation for Multi-channel Drop-shippers Annibal Sodero, PhD, University of Arkansas Drop-shipping, an arragement in which retailers sell products online and the products are sent directly from the vendors’ facilities to the point of consumption, is growing in importance. For vendors of seasonal products who drop-ship but also sell wholesale to their retail customers, there are many questions left regarding inventory allocation and deployment, because they need to consider the usual tradeoffs between inventory costs and service levels. For instance, when should they segregate or keep a single pool of inventory to satisfy demand? When and whom to release inventory in the wholesale channel? We simulate scenarios to provide answers to those questions, among others. Arizona State University, Tempe, AZ, United States, Elliot Rabinovich, Sungho Park, Minha Hwang Warehouse club (WC) retailers have historically relied on low-density store networks as a source of competitive advantage. Despite their significant presence

n WB07 North Bldg 123 Models and Methods for Network Design Sponsored: Optimization/Network Optimization Sponsored Session

Chair: Angelika Leskovskaya, Southern Methodist University, Caruth Hall 3145 Dyer Street, Suite 372, Dallas, TX, 75275, United States 1 - Robust Service Network Design under Demand Uncertainty: An Exact Approach

Mingyao Qi, Associate Professor, Tsinghua University, Room E204A, Shenzhen, 518055, China, Zujian Wang

Freight forwarding companies are faced with difficulties in handling uncertainties, especially demand uncertainty under the circumstances of no enough historical data or accurate forecasting approach. In this paper, we propose a two-stage robust optimization method for service network design under demand uncertainty. We employ probability-free uncertainty sets to describe the possible scenarios, and implement a column-and-constraint generation algorithm to solve the proposed two-stage robust models exactly and efficiently. Numerical results indicate that the proposed algorithm outperforms the Benders decomposition method in both solution quality and computational time. 2 - Service Network Design with Node Capacity with Application in Parcel Delivery Huan Jin, MIT Global SCALE Network, 462 Wenyuan Road, Bldg 3, Rm 504, Ningbo, 315100, China The rapid development of e-commerce has made the explosion of online shopping. A large part of the urban logistics parcels come from online e- commerce orders. We consider a service network design for parcels urban logistics with node capacity. The node capacity is the space/quantity limit of holding resources (drivers, vehicles) at terminals. The node capacity, for instance, the parking limit at terminals, varies by the time of a day or the day of a week. We introduce a general formulation and design a Branch-and-Price method with heuristic algorithm to provide good quality solution of the formulation. A comprehensive analysis of the formulation and algorithm is given based on the numerical results. 3 - Integer Programming Models for Freight Logistics Service Network Design with Tree Constraints Less-than-truckload (LTL) service providers often, in practice, route shipments so as to satisfy tree constraints: at any node in their service network, all shipments that are handled at that node and that have the same destination must be routed via the same outgoing arc. Although the design of LTL service networks has been widely studied, tree constraints have generally been ignored. Here, we consider integer programming (IP) models for the service network design problem with tree constraints. We formulate and analyze multiple IP models and report computational results. 4 - A Mathematical Programming Model with Equilibrium Constraints for Competitive Closed Loop Supply Chain Network Design Qiang Qiang, Penn State University, Management Division, 30 E. Swedesford Rd, Malvern, PA, 19355, United States, Yuxiang Yang, zuqing huang A firm sets up his facilities including manufacturing/remanufacturing plants and distribution/collection centers, incorporating an existing CLSC network. The entering firm has to compete with the existing firms in the existing network. The entering firm behaves as the leader of a Stackelberg game while the existing firms in the existing network are followers. We use a CLSC network equilibrium model to capture the existing firms’ reaction. A mathematical programming model with equilibrium constraints is developed for this competitive CLSC network design problem. Numerical examples and the related results are studied for illustration purpose. 5 - Data-driven Transit Network Design Julia Y. Yan, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States, Dimitris Bertsimas, Yee Sian Ng Modern transit agencies have to provide reliable and equitable transit services in the face of declining ridership and tight financial budgets. In 2012, this analysis was performed for MBTA bus services, involving a tedious procedure of “manual branch-and-boundö. Instead, we provide a unified approach that optimally selects a subset of the routes to maximize ridership, while satisfying coverage requirements. It solves within minutes for networks with hundreds of services across 164 municipalities. In comparisons against a greedy heuristic, our approach is able to double the transit ridership while satisfying regulatory requirements under tight financial constraints. Ira M. Wheaton, Postdoctoral Fellow, Georgia Institute of Technology, 755 Ferst Drive, NW, Atlanta, GA, 30332, United States, Natashia Boland

442