2016 INFORMS Annual Meeting Program

WC27

INFORMS Nashville – 2016

2 - Developing A Non Invasive Tool To Diagnose Diabetic Retinopathy Saeed Piri, PhD Student, Oklahoma State University, School of Industrial Engineering & Management, Engineering North, Stillwater, OK, 74078, United States, saeed.piri@okstate.edu, Tieming Liu, Dursun Delen Diabetic retinopathy is the leading cause of vision loss in the US. In this study a non-invasive tool is developed to detect this disease. The demographic and lab data of more than 1.4 million diabetics has been analyzed and four sets of predictive models have been developed. The first set encompasses the models that have been developed using lab and demographic data. In the second set, comorbidity data was included in addition to basic data. Third set consists of models that are built using the oversampled data by applying SMOTE method. Fourth set includes ensemble models that have been developed using the outputs of different single predictive models. The accuracy of the best model is close to 90%. 3 - Call For Bids To Improve Matching Efficiency: Evidence From Online Labor Markets Xue Guo, Doctoral Student, Temple University, Fox School of Business, 1801 North Broad Street, Philadelphia, PA, 19122, United States, xueguo@temple.edu, Jing Gong, Pavlou Paul In online labor markets, Call for Bids (CFB) plays an important role in affecting the efficiency and quality of labor matches. In this study, we analyze the text of CFBs and examine five attributes associated with the clarity of CFBs that may affect the contracting success rate and post-project satisfaction—codifiability, uncertainty, complexity, flexibility and monitoring. We propose to use deep learning algorithms to analyze large unstructured textual data and test our model using archival data from one of the largest online labor platforms. 4 - An Empirical Investigation Of Online Marketplace Users Yuanyuan Shen, Stanford University, Stanford, CA, 94305, United States, yyshen@stanford.edu, Haim Mendelson, Kenneth Moon We develop a simple Markovian model of user behavior which incorporates user learning and social value on an online marketplace. We test the model and analyze additional tradeoffs and market characteristic. We use data from Kiva, a non-profit lending marketplace. Consistent with the predictions of our model, we find evidence of user engagement begetting greater engagement. Meanwhile, an engaged user’s tenure is negatively correlated with his lending rate while the correlation turns to be positive when the user has no loans. We also identify significant differences in the behaviors of user acquired directly and those acquired virally. WC27 201A-MCC Innovations in Retail Operations Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Nicole DeHoratius, University of Chicago, 58071 S Woodlawn Ave., Chicago, IL, 60637, United States, Nicole.DeHoratius@ChicagoBooth.edu 1 - Retail Inventory Liquidation Value: Local Economic Factors Nathan Craig, OSU, craig.186@osu.edu The liquidation value of a firm’s assets is an input to numerous models. Our research characterizes liquidation values in practice and identifies factors that affect liquidation value. In the context of retail inventories, we analyze data from multiple sources to identify economic features of a store’s local market that significantly impact inventory liquidation value. 2 - Here Comes The Sun: Measuring And Exploiting Weather Shocks In Fashion Retailing Victor Martinez-de-Albeniz, Professor, IESE Business School, Barcelona, 08034, Spain, valbeniz@iese.edu, Abdel Belkaid We model and empirically study the impact of weather variables on the operations of a large apparel retailer. Specifically, we focus on traffic store and the conversion of traffic into product category sales. We find that rain increases traffic but decreases conversion in shopping mall stores, while the opposite is true in street stores. Lower temperatures increase traffic and increase conversion of sales of the “appropriate” categories (summer vs. winter). Finally, we demonstrate how to exploit weather variations to improve profits. 3 - Optimizing Prepacks To Ship Products In Retail Supply Chains Stephen A Smith, Santa Clara University, Omis Lucas Hall 216h, 500 El Camino Real, Santa Clara, CA, 95053-0382, United States, ssmith@scu.edu, Naren Agrawal We present recent results from the analysis of shipping and inventory policies in retail supply chains when multiple pre-packs are permissible. We compute exact solutions to a dynamic stochastic programming formulation and derive insights based on numerical analysis of a sample case study.

4 - Package Size And Pricing Decisions With A Bulk Sale Option Ismail Kirci, University at Texas at Dallas, ismail.kirci@utdallas.edu We investigate the package size and pricing decisions of a retailer selling a perishable product that can be offered in packages or in a container that allows consumers to buy as much as they want (Bulk Sale). We show how the existence of a bulk sale option affects the optimal package size, the optimal price and product waste at the consumer level. WC29 202A-MCC Mechanisms to Enhance the Value of Used and Returned Products Sponsored: Manufacturing & Service Oper Mgmt, Sustainable Operations Sponsored Session Chair: Eda Kemahlioglu-Ziya, Poole College of Management at North Carolina State University, 2801 Founders Dr, Raleigh, NC, 27695, United States, ekemahl@ncsu.edu 1 - Impact Of Recycling Standards On E-waste Recovery In The Presence Of Secondary Market Gokce Esenduran, Ohio State University, esenduran.1@osu.edu, Yen-Ting (Daniel) Lin, Wenli Xiao, Minyue Lin There are two voluntary certification programs for e-waste recyclers: e-Stewards and R2. While both provide assurance of proper e-waste handling, e-Stewards is much stricter than R2, resulting in a higher processing cost. With the aim of identifying when a recycler should adopt e-Stewards versus R2, we model the competition between two recovery channels; each consisting of a recycler and a collector who can sell collected e-waste either in the secondary market or to its recycler. We find that devoid of competition, a recycler always chooses R2; whereas under competition, recyclers may choose either one. We also examine how secondary market and scale economies affect recyclers’ decisions. 2 - Design For Reusability And Product Reuse Under Radical Innovation Michael Galbreth, University of South Carolina, galbreth@moore.sc.edu Many industries, including consumer electronics and telecommunications equipment, are characterized by short product lifecycles, constant technological innovations, rapid product introductions, and fast obsolescence. Firms in such industries need to make frequent design changes to incorporate innovations, and the effort to keep up with the rate of technological change often leaves little room for the consideration of product reuse. In this paper, we study the design for reusability and product reuse decisions in the presence of both a known rate of incremental innovations and a stochastic rate of radical innovations over time. 3 - Extracting Maximum Value From Consumer Returns Cerag Pince, Kuhne Logistics University, Cerag.Pince@the-klu.org, Mark Ferguson, Beril L Toktay Consumer returns constitute a substantial fraction of sales in the consumer electronics industry and often cannot be re-sold as new due to litigation concerns. Therefore, identifying the best joint pricing and disposition strategy is a challenging but important decision for consumer electronics OEMs. This paper investigates how an OEM should price new and refurbished products while allocating consumer returns between remarketing and warranty coverage options over the product’s short life cycle. 4 - Truth-inducing Mechanism For Medical Surplus Products Allocation Can Zhang, Georgia Institute of Technology, Atlanta, GA, 30309, United States, czhang2012@gatech.edu, Atalay Atasu, Turgay Ayer, Beril L Toktay We study a product allocation problem faced by a Medical Surplus Recovery Organization (MSRO) that recovers and manages reusable medical products to fulfill the needs of under-served healthcare facilities in developing countries. We focus on designing truthful mechanisms to elicit the recipients’ needs information, and we prove that the optimal truthful mechanism has a very simple structure. We further show that our mechanism significantly improves the MSRO’s total value provision compared with the current practice and previously proposed recipient-driven models.

436