2015 Informs Annual Meeting

WA57

INFORMS Philadelphia – 2015

WA55 55-Room 108B, CC Forecasting Contributed Session Chair: Arnd Huchzermeier, WHU-Otto Beisheim School of Management, Production Management, Burgplatz 2, Vallendar, 56179, Germany, ah@whu.edu 1 - Modeling Effects of Price Changes on Sales for Improved Forecasting and Promotion Planning Shubhankar Ray, Walmart Labs, 850 Cherry Ave., San Bruno, CA, 94066, United States of America, sray02@walmartlabs.com, Brian Seaman, Ashin Mukherjee In online retail, the high sales variability & missing samples makes item-level estimation of price elasticities unreliable. We use a mixed-effects framework to fit autoregressive distributed lag models on item groups to get reliable population- level estimates of short/long-run elasticities while pooling information across items. This model can be layered with any forecasting model to generate price sensitive forecasts & can help steer the sales in a desired direction during promotion planning. 2 - A Better Accuracy Measure than Mean Absolute Percentage Error Chris Tofallis, Professor of Decison Science, Hertfordshire Business School, College Lane, Hatfield, AL10 9AB, United Kingdom, c.tofallis@herts.ac.uk Although Mean Absolute Percentage Error is widely used, it suffers from bias: its use in selecting a forecasting method favours methods whose predictions are too low. We explain why this happens and present an alternative relative accuracy measure which is unbiased. This measure can be used to estimate prediction models. Minimum MAPE models do not predict a known statistic. By contrast, when the proposed metric is used, the resulting least squares regression model predicts the geometric mean. 3 - Information Sharing in Supply Chains Facing Autoregressive Demand. We introduce a general class of potentially valuable sharing arrangements in a multi-stage supply chain in which the retailer observes stationary ARMA demand with respect to Gaussian white noise (shocks). We demonstrate how a typical supply chain player can create an incentive-compatible sequence of partial- information shocks (PIS) based on its available information and share these with an adjacent upstream player, potentially reducing the upstream players lead-time demand forecast error. 4 - Judgemental Demand Forecasts for Online Sales of a Premium Bike Manufacturer Arnd Huchzermeier, WHU-Otto Beisheim School of Management, Production Management, Burgplatz 2, Vallendar, 56179, Germany, ah@whu.edu, Christoph Diermann We present a judgmental demand forecast model for a premium bike manufacturer selling each season a new collection of products online. Forecast accuracy could be improved by using a team approach, by de-biasing the expert forecasts and by selecting the appropriate segmentation of bike types. Forecast accuracy could be improved by 26% and profit increased by more than 40%. WA56 56-Room 109A, CC Operations Management VII Contributed Session Chair: Xueyuan Cai, Huazhong University of Science and Technology, 1037Luoyu road, Wuhan, China, comeon1644@163.com 1 - Configurations of Distribution Strategies Jing Tang, Em-Lyon Business School, 23 Avenue Guy de Collongueresear, Ecully, 69134, France, tangacade@gmail.com, Yeming Gong Based on 124 quantitative samples with both first-hand and second hand data, as well as 56 qualitative samples, this paper examines the strategic fit of distribution strategies from the perspective of configuration theory. We find that the fit between operational decisions including infrastructural and structural decisions, and operational competencies including cost and flexibility, has an important effect on business performance. Vladimir Kovtun, Syms School of Business, 215 Lexington Avenue, New York, NY, 10016, United States of America, vladimir.kovtun@yu.edu, Avi Giloni, Clifford Hurvich

2 - Exchange-old-for-new Program: An Incentive to Induce Early Purchases with Product Rollover Yongbo Xiao, Associate Professor, Tsinghua University, School of Economics and Management, Beijing, 100084, China, xiaoyb@sem.tsinghua.edu.cn, Qian Liu Many customers may choose to wait for new generation of products in face of product roll-overs. We propose to adopt an exchange-old-for-new program to induce customers to make early purchase. Starting from the choice behavior of strategic customers and based on a dynamic programming model, we study the optimal pricing decision involved in the exchange program. 3 - Model and Algorithms for the Integrated Production and Vehicle Routing Problem In this paper, the integrated production and vehicle routing problem is considered in a Make-to-Order (MTO) manufacturer, where there is a single machine for production and limited vehicles with capacity constraints for transportation. The objective is to determine the decisions about production scheduling, transportation batching and vehicle routing, to minimize the maximum order delivery time. An optimal property is proposed and based on the property,an improved genetic algorithm is developed. 4 - Optimizing Multi-item Joint Replenishment under Stochastic Demand and Uncertain Leadtime Xueyi Ai, School of Management, Huazhong University of Science and Technology, 1037 Luoyu Road, Wuhan, China, Wuhan, China, aixueyi1030@gmail.com, Jinlong Zhang, Lin Wang In real-life practice, leadtime uncertainty is a common phenomenon which may be caused by various factors. The purpose of this paper is to develop a more realistic joint replenishment model considering both stochastic demand and uncertain leadtime simultaneously, and propose an improved fruit fly optimization algorithm (IFOA) to solve the problem, minimizing the long-run average total costs. Experimentation results show the standard deviation of leadtime has large effect on the system. 5 - Strategy Comparison of Capacity Allocation in a Supply Chain with One Dominant Retailer Xueyuan Cai, Huazhong University of Science and Technology, 1037Luoyu road, Wuhan, China, comeon1644@163.com, Jianbin Li We consider a distribution system with one supplier and two retailers. The supplier will implement allocation mechanisms or improve her wholesale price when the capacity can’t match demand. Conventional wisdom thinks that the impact to the supplier by “order-inflation” mechanisms is not the same as that by “truth-telling” mechanisms. But we show that this is not true in our setting where each retailer faces a oligopoly market and places order to the supplier considering the capacity constraints. WA57 57-Room 109B, CC Electric Transportation Systems Modelling Sponsor: ENRE – Energy I – Electricity Sponsored Session Chair: Ramteen Sioshansi, Associate Professor, The Ohio State University, Integrated Systems Engineering, 1971 Neil Avenue, Columbus, OH, 43210, United States of America, sioshansi.1@osu.edu 1 - Optimization of Incentive Polices for Plug-in Electric Vehicles Yu Nie, Northwestern University, Evanston, IL, United States of American, y-nie@northwestern.edu, Mehrnaz Ghamami High initial purchase prices and the lack of supporting infrastructure are major hurdles to the adoption of plug-in electric vehicles (PEVs). It is widely recognized that the government could help break these barriers through incentive policies, such as offering purchase rebates and funding charging stations. The objective of this paper is to propose a modeling framework to optimize the design of such incentive policies. 2 - Incentive-compatible Charging Mechanisms for Plug-in Electric Vehicles Mark Nejad, Assistant Professor, University of Oklahoma, Industrial and Systems Engineering, Norman, OK, United States of America, mark.nejad@ou.edu, Lena Mashayekhy We design charging mechanisms for plug-in electric vehicles considering different charging rates. We prove that our proposed mechanisms are dominant-strategy incentive compatible, so that EV drivers have no incentive to misreport their charging requests or their arrival-departure dynamics. Ling Liu, Huazhong University of Science & Technology, 1037 Luoyu Road, Wuhan, Hubei, China, Wuhan, China, 182028870@qq.com

391