Informs Annual Meeting 2017

POSTER SESSION

INFORMS Houston – 2017

33 - Cloud-based Inventory Optimization Solution for Retail Dmitry Pechyony, Microsoft, Cambridge, MA, United States, dmpechyo@microsoft.com, Chenhui Hu, Hong Lu, Praneet Singh Solanki, Ilan Reiter We present a cloud-based inventory optimization solution. It can be deployed in Azure cloud from https://aka.ms/invopt and customized to business goals and constraints of retailers. Our solution has a flexible, scalable and extensible architecture. Operations researchers and developers can use it to define optimization problems of new inventory management policies, operationalize them and track their performance. We use Azure Data Lake to store unlimited data volumes and Azure Batch to solve optimization problems in parallel. Furthermore, many free and commercial solvers can be plugged into our solution using Pyomo interface. The source code and documentation are at https://aka.ms/invdoc . 34 - Information Overload’s Hidden Impact – the Response to a Rising Tide of Information. William D. Senn, Dr., Texas Woman’s University, 8808 Cassidy Lane, Aubrey, TX, 76227, United States, wsenn1@twu.edu As information load increases, decision accuracy also increases up to an undefined point after which decision accuracy decreases. This turning point is where information overload begins, according to (Eppler & Mengis, 2004). This study looks at how knowledge managers perceive information overload. It investigates the impact of fake news and big data on these perceptions. In a world of ever increasing information flow people are bombarded with information. This research provides insight into the information overload response function. 35 - Ranked Constraint Reconciliation for Price Optimization Jagdish Ramakrishnan, Data Scientist, Walmart Labs, 1000 National Ave, Unit 214, San Bruno, CA, 94066, United States, jagdish.ram@gmail.com, Matyas Sustik Given a set of ranked infeasible constraints, generate a new feasible set of constraints. In the context of price constraint reconciliation, merchants provide local constraint information (e.g., item price bounds, pack sizes) that together are impossible to meet. We would like to avoid violations on high priority constraints and allow some violations on low priority constraints. The proposed method processes constraints by rank and repeatedly solves optimization subproblems using slack variables. We implemented a python module that takes in ranked linear constraints and outputs a set of feasible constraints. It enables maintainability and flexibility compared to lengthy logic-based code. 36 - Biological Restoration of Fragmented Landscapes for Connectivity and Intactness Denys Yemshanov, Research Scientist, Natural Resources Canada, 1219 Queen street East, Sault Ste. Marie, ON, P6A2E5, Canada, denys.yemshanov@canada.ca, Robert G. Haight, Frank H. Koch, Maintaining habitat connectivity for endangered wildlife is critical for successful restoration of fragmented landscapes. We propose a maximum flow, multigraph connectivity formulation with connected area requirements and budget constraint to address landscape restoration. The MIP model is scenario-based, with an objective of determining restoration strategies that maximize habitat area accessible to wildlife in a fragmented landscape. We apply the model to habitat restoration for woodland caribou in boreal forest in Cold Lake area, Alberta, Canada. 37 - Investigating the Effect of Prescription Sequence on Developing Adverse Drug Reactions Adverse Drug Reactions (ADR) is a pharmacovigillance term that refers to any injury caused by taking drugs in the way prescribed by physicians. In this study applying an emergent pattern mining approach to electronic health records of more than 370,000 diabetic patients, we examine the potential effect of drug prescription sequence on the development of ADRs. Specifically we focus on Acute Renal Failure as a condition that frequently occurs in diabetic patients. We first investigate whether common diabetic drugs play any confounding role in developing renal failure; afterwards, the effect of prescription sequence on this ADR is examined. The results verify that such role exists. 38- Analysis on the Coordinated Development Between Logistics Industry and the Other Industries in Sichuan Based on Grey Relativity Analysis Si Chen, Dr., Southwest Jiaotong University, #111 North of Erhuan Road, ChengDu, 610031, China, chensi@swjtu.edu.cn, Xinyuan Li, Yinying Tang, Mi Gan Logistics industry plays an important role in the development of Sichuan economy. We research into the relationship between the logistics industry and the agriculture,manufactory and business, which are the main economic development departments in Sichuan. And we analysis the gray correlation and the coupling degree of logistics industry and the other industries based on the logistics related data of Sichuan province from 2011 to 2016. Therefore, the development of logistic industry is a strong influence to the other industries, and it can provide adaptive supporting environment forthe other industries. Behrooz Davazdahemami, Oklahoma State University, 4599 N. Washington St, Apt 40A, Stillwater, OK, 74075, United States, davazda@okstate.edu, Dursun Delen Marc-André Parisien, Tom Swystun, Quinn Barber, Salimur Choudhury, Fabio Campioni, Cole Burton

39 - Pricing and Service Strategies in E-retailing: the Effects of Safety Stock Yi Ding, Southeast University, Sipailou 2, Nanjing, 210096, China, drdingyi@outlook.com This study aims to examine service and price competition in an online retailing system comprised of two suppliers and one e-retailer. A Stackelberg game is formulated with the suppliers as the leaders who determine wholesale prices and service times and the e-retailer as the follower who sets the retail prices. Changes of service time affect safety stock which is characterized by the guaranteed service model. We find that the “service first” strategy should be pursued when suppliers face increasing safety stock cost, and the influence of one supplier’s service can be spilled over to the competing product’s wholesale and retail prices, depending on the difference between consumer transfer ratios. 40 - Forecasting Strategies for Predicting the Peak Load Demands of a University Campus: A Case Study on Rochester Institute of Technology Academic institutions spend thousands of dollars every month on their electric power consumption. A large percentage of these institutions follow a demand charges pricing structure; here the amount a customer pays to the utility not just depends upon the energy consumed during the month but also on the maximum load required by the customer over a moving window of time as decided by the utility. This research is a case study conducted on Rochester Institute of Technology which aims at developing a forecasting strategy to predict the occurrence of peak load periods and hence run a demand response plan to maximize the financial savings for the institution. 41 - Signal Effect of XBRL Filing: An Investor’s Perspective Soohyun Cho, Assistant Professor, Rutgers University, Newark, NJ, United States, scho@business.rutgers.edu The SEC requires companies to provide financial data in the XBRL format as well as traditional financial statements. Despite XBRL disclosures containing information identical to the statements, and belying corporate sector doubt of the disclosures’ benefits, most capital market investors tend to respond positively to the filing. For market reaction analysis, we develop a stylized model using two investor groups - market leaders and market followers, depending on their respective information sets. Applying this model, we find that the level of the XBRL filing’s capital market impact is associated with the number of market followers. 42 - An Analysis of One Version of the Tracking Signal Peruvemba Sundaram Ravi, Associate Professor, Wilfrid Laurier University, Operations & Decision Sciences, Waterloo, ON, N2L.3C5, Canada, pravi@wlu.ca Tracking signals are used to determine whether a sequence of forecasts is unbiased. We analyse a version of the CUSUM tracking signal that is used in standard Operations Management texts. We show that this version of the tracking signal can indicate that the sequence is not unbiased even when it is unbiased. We suggest a simple modification that serves to eliminate this flaw. 43 - Data-driven Decision Making for Re-timing and Optimization of Signalized Intersections Nabaruna Karmakar, PhD Candidate, North Carolina State University, 3609 Research Building IV, Raleigh, NC, 27695, United States, nkarmak@ncsu.edu, Thomas Chase, Billy M. Williams This research is aimed at using a variety of data sets to enable intelligent decision- making regarding signal re-timing and maintenance. The final product of this research is building a data-driven optimization technique, using mathematical programming to identify corridors that need re-timing and suggesting different time of day signal timing plans for such corridors. With continuous data like high resolution controller data and Vehicle Probe Data, enhanced performance metrics can be calculated to strategically prioritize and select systems for re-timing that have a higher potential of improvement solely based on optimization benefits. 44 - Content Discovery and User Engagement in Digital Platforms: A Field Experiment in Music Platforms Shadi Janansefat, PhD Candidate, University of Pittsburgh, 229 Mervis Hall, Pittsburgh, PA, 15260, United States, shj35@pitt.edu Digital content provider platforms focus on improving success metrics such as engaged time and content discovery, which create continuous usage and engagement. We draw from information foraging theory and social foraging model to investigate how platforms can use social design to improve users’ engagement and content discovery. We employ a randomized field experiment on a social music platform mobile application to study how availability of two information signals, popularity on platform and popularity among peers, can enhance user engagement and content discovery. We enumerate the experiment design to test our hypotheses. Harshit Saxena, Master of Science, Rochester Institute of Technology, 1 Lomb memorial drive, Rochester, NY, 14623, United States, hs3185@rit.edu, Katie McConky

198