Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WB16

5 - Optimal Staffing under Endogenous Arrivals with Heterogeneous Customer Time-of-service Preferences Yang Li, Chinese University of Hong Kong, 12 Chak Cheung Street, Cheng Yu Tung Building, Hong Kong, Hong Kong, Philipp Afeche The service operations literature usually treats arrivals as exogenous processes. However, arrival processes may be endogenous in many settings. That is, customers may account for system congestion in choosing their time of service. We propose an equilibrium model that captures how rational customers with heterogeneous preferences decide their time-of-service. We also study the optimal staffing policies, taking into account customers’ time-of-service choices. n WB16 North Bldg 127B Transitions toward Sustainable Energy, Food, and Living Sponsored: Manufacturing & Service Oper Mgmt/Sustainable Operations Sponsored Session Chair: Owen Wu, Indiana University, Bloomington, IN, 47405, United States Co-Chair: Yangfang Zhou, Singapore Management University, Singapore, 178899, Singapore 1 - Kicking Ash: Who (or What) is Winning the War on Coal David F. Drake, Harvard Business School, Morgan Hall 425, Boston, MA, 02163, United States, Jeffrey York Power generators throughout the U.S. have shed coal capacity at an unprecedented rate over the past several years. Multiple stakeholders have claimed credit - natural gas executives, policy makers, renewables advocates, and environmental NGOs. Through a survival analysis, we explore the extent to which each has impacted the expected life of coal-fired power generating units. 2 - Greenest Grocer: Online or Offline? Ekaterina Astashkina, INSEAD, Boulevard de Constance, Fontainebleau, 77305, France, Elena Belavina, Simone Marinesi We compare environmental differences of traditional and online grocery retail channels. We build and calibrate the model that captures household food-buying patterns and the shopping mode choice. We find that, in high store density cities, the development of an online outlet is more efficient in reducing environmental impact in contrast to increased store density. In particular, online shopping attracts households for whom offline is least convenient — such households have the worst impact: they build up more inventories and travel longer distances. Higher store density, in turn, impacts households in a less targeted fashion and, Fariba Farajbakhsh Mamaghani, University of Texas at Dallas, 800 West Campbell Road SM 30, Richardson, TX, 75080, United States, Saed Alizamir, Shouqiang Wang Matching demand with supply has been a key challenge in operating residential electricity markets. Existence of exogenous random shocks (e.g., outdoor weather condition) and consumer’s limited capability in adjusting their household appliance’s settings on the other hand lead utility firms face stochastic demand functions which is not well understood in the literature .In this paper, we construct a demand model to explicitly account for such limited consumer response to changes in exogenous random shocks and We fully characterize the firm’s optimal price. 4 - Plugged in at the Right Time? Managing Electrical Vehicles Charging As electric vehicles (EV) increasingly penetrate the electrical grid, new business models emerge to reduce the cost of charging EV batteries. We analyze how cost savings depend on the time windows that EVs are plugged in and the charging amount required by customers. We consider setting appropriate prices for the EV charging, in order to change customers’ charging demand to maximize the cost savings. Owen Wu, Indiana University, Kelley School of Business, 1309 E. 10th Street, Bloomington, IN, 47405, United States thus, is more efficient only in very low store density scenarios. 3 - Electricity Pricing with Limited Consumer Response

n WB17 North Bldg 127C Data-Driven Supply Chain Strategies Sponsored: Manufacturing & Service Oper Mgmt/Supply Chain Sponsored Session Chair: Yun Zhang, Massachusetts Institute of Technology, 77 Massachusetts Ave, E18, Cambridge, MA, 02139, United States 1 - Value of Analytics in Inventory Systems Li Wang, Massachusetts Institute of Technology, 77 Massachuestts Ave, Cambridge, MA, 02139, United States, Jussi Keppo, David Simchi-Levi We study the trade-off between holding safety stock and deploying data analytics for a retailer facing demand uncertainty. In Newsvendor models, safety stock is used to mitigate the risk of stockouts due to demand uncertainty. However, the demand risk can be reduces by investing in data analytics that creates more accurate demand forecasts and this way lowers the need of safety stock. We solve this joint problem of optimal quantity of analytics and safety stock, and study the comparative statics of the model. For instance, we show that the lower the cost of analytics is, the lower the safety stock. 2 - Distributionally Robust Dynamic Inventory Management with Bank Finance Zhenzhen Yan, Assistant Professor, Nanyang Technological University, 21 Nanyang Link, Singapore, 637371, Singapore, Ruijie Zhang, Yuanguang Zhong In this paper, we study a stochastic dynamic inventory scheduling problem faced by a capital-constrained company who periodically replenishes its inventory from a supplier. The problem is to determine the optimal inventory and financing decisions to maximize its expected terminal cash flow. Instead of assuming a certain demand distribution, we only require mean and variance of demand are known. We formulate this problem as a convex conic optimization problem which can be approximated using a positive semi-definite program. We study and compare two types of financing strategies: unsecured loans and asset-based loans. 3 - Quantile Forecasting and Data-driven Inventory Management under Nonstationary Demand Ying Cao, University of California, Berkeley, CA, 94720, United States, Max Shen, Max Shen Most work in data-driven inventory management assume that historical demand can be regarded as samples drawn from a real distribution; or when demand distribution changes over time, multiple sample paths are available. In reality, however, demand is time-correlated and may exhibit nonstationarity. In this talk, we consider a general autoregressive demand process without prespecified parametric structure. Based on which, we present a neural network framework for predicting its quantiles, and argue that it is a data-driven approach for determining stock levels in the environment of newsvendor problem and its multiperiod extension. This is a joint work with Professor Max Shen from UC Berkeley. 4 - Supply Chain Network Design and Coordination Yun Zhang, Massachusetts Institute of Technology, Cambridge, MA, United States, David Simchi-Levi We explore network design and coordination strategies in the context of modern supply chains, with a focus in public health. n WB18 North Bldg 128A Operations Management II Contributed Session Chair: Xinping Wang, College of Engineering, Nanjing Agricultural University, 40 Dianjiangtai Road, Nanjing, 210031, China 1 - The Effect of Merger and Acquisition on Inventory Turnover Zhihao Zhang, Student, University of South Carolina, Columbia, SC, United States We evaluate the effect of merger and acquisition (M&A) on firms` inventory turnover, a well-developed measure of operations performance. By applying System Generalized Methods of Moments (GMM) and Discontinues Growth Modeling (DGM), we are able to empirically test and show that mergers and acquisitions interrupt companies` inventory turnover in the short run, but eventually benefits companies in the long run. We also find that companies who face high demand uncertainty M&A benefits the most from the long-term synergies they created through M&A.

446