Informs Annual Meeting 2017

TD17

INFORMS Houston – 2017

TD17

optimality are only for single-product systems. We have formulated a multi-stage Stochastic Program (SP) and shown its optimal solutions coordinate inventory positions for minimizing inventory cost in multi-product ATO systems with a general Bill of Materials. We will discuss properties of SP solutions that allow their implementation in dynamic ATO systems. 2 - Learning Algorithms in Inventory and Pricing Models

340A Service Queues Sponsored: Applied Probability Sponsored Session Chair: Rouba Ibrahim, University College London, University College London, London, WC1E 6BT, United Kingdom, rouba.ibrahim@ucl.ac.uk 1 - Choice Models, Queueing Theory, and Delayed Information Jamol Pender, Cornell University, Ithaca, NY, United States, jjp274@cornell.edu Many service systems provide queue length information to customers thereby allowing customers to choose among many options of service. However, queue length information is often delayed and is often not provided in real time. Recent work by Dong et al. explores the impact of these delays in an empirical study in U.S. hospitals. In this talk, we prove an appropriately scaled queueing model with customer choice and constant delays converges to a delay differential equation. We analyze these delay differential equations and derive the exact threshold that determines whether the queues have asynchronous dynamics. 2 - Maximizing Ridership in Bicycle Sharing Systems using Empirical Data and Stochastic Models Pradeep Kumar Pendem, University of North Carolina, Chapel Hill, 110 Mistywood Cir, Apt S, Chapel Hill, NC, 27514, United States, pradeep.pendem@gmail.com, Vinayak V. Deshpande We analyze the optimal allocation of bikes in a network of stations to improve ridership and service level under non-stationarity demand and station substitution. Our stochastic demand model captures both bike pickups and drop- offs, as well as demand non-stationarity and substitution under stockout. We utilize large datasets on trips, real time inventory information at stations, distances between stations in combination with the model to estimate demand parameters. The optimal allocation of bikes across stations to maximize ridership is determined using a dynamic program. The optimal policy informed by our model could improve ridership and service level by 7.60% and 1.69% to current policy. 3 - Managing Customer Expectations and Priorities in Service Systems Gad Allon, The Wharton School, Philadelphia, PA, United States, gadallon@wharton.upenn.edu, Qiuping Yu, Achal Bassamboo We study how to use delay announcements to manage customer expectations while allowing the firm to prioritize among heterogeneous customers. When the firm has information about the state of the system, yet lacks information on customer types, delay announcements play a dual role: they inform customers about the state of the system, while they also have the potential to elicit information on customer types based on their response to the announcements. The tension between these two goals has implications to the type of information that can be shared credibly. 4 - Announcing Delays when the Present Depends on the Future Rouba Ibrahim, University College London, MS& I.department, UCL, Gower Street, London, WC1E 6BT, United Kingdom, rouba.ibrahim@ucl.ac.uk, Achal Bassamboo We investigate the accuracy of announcing the waiting time of the Last customer to Enter Service (LES) in a queueing model with multiple customer classes and a priority service discipline. We present ways of exploiting this historical information to design new and improved announcements and supplement our theoretical results with an empirical study. 340B Joint Session MSOM/APS: Supply Chain and Inventory in Applied Probability I Sponsored: Applied Probability Sponsored Session Chair: Linwei Xin, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, United States, lxin@illinois.edu 1 - Using Multi-stage Stochastic Programs for Inventory Position Coordination in Assemble-to-order Inventory Systems Qiong Wang, University of Illinois at Urbana-Champaign, Urbana-Champaign, IL, United States, qwang04@illinois.edu, Marty Reiman, Haohua Wan To manage Assemble-to-Order (ATO) inventory systems efficiently, the inventory positions of components should be aligned with the availability of other components with longer lead times. Independent base stock policies do not feature such a coordination mechanism and existing schemes with proven TD18

Cong Shi, University of Michigan, 2257 Woodhaven Court, Ann Arbor, MI, 48105, United States, shicong@gmail.com, Xiuli Chao, Boxiao Chen, Huanan Zhang

In recent years, online retailing firms have been experimenting and implementing innovative dynamic pricing strategies and inventory policies to better match demand with supply. In many cases, the decision maker may not know the demand distributional information of a given product a priori, and can only rely on sales data. We propose nonparametric learning algorithms for several fundamental inventory and pricing models with unknown demand functions under censored demand information. We show that the proposed algorithms converge to the clairvoyant optimal policies as the planning horizon increases, and obtain the convergence rate of regret. 3 - Dynamic Pricing and Inventory Control with Large Replenishment Lead Times Linwei Xin, University of Chicago, 5807 S Woodlawn Avenue, Chicago, IL, 60637, United States, Linwei.Xin@chicagobooth.edu, Xin Chen, Yuanling Gan We consider a joint pricing and inventory control problem with positive replenishment lead times. Although this problem has been extensively studied in the literature, the structure of the optimal policy remains poorly understood. In this work, we propose a class of so-called constant-order list-price policies. We prove that the best constant-order list-price policy is indeed asymptotically optimal as the lead time grows large, which is exactly the setting in which the problem becomes computationally intractable due to the curse of dimensionality. We also show that the best constant-order list-price policy can be computed effectively. 342A Wind Power Sponsored: Energy, Natural Res & the Environment Environment & Sustainability Sponsored Session Chair: Amelia Musselman, Georgia Institute of Technology, Atlanta, GA, 30318, United States, amusselman@gatech.edu 1 - Integrated Predictive Analytics & Optimization for Opportunistic Maintenance and Operations in Wind Farms Murat Yildirim, Georgia Institute of Technology, 765 Ferst Drive NW, # 415, Atlanta, GA, 30332, United States, murat@gatech.edu, Nagi Gebraeel, Andy Sun We propose an integrated framework for wind farm maintenance that combines i) predictive analytics methodology that uses real-time sensor data to predict future degradation and remaining lifetime of wind turbines, with ii) a novel optimization model that transforms these predictions into profit-optimal maintenance and operational decisions for wind farms. To date, most applications of predictive analytics focus on single turbine systems. In contrast, this presentation provides a seamless integration of the predictive analytics with decision making for a fleet of wind turbines scattered across multiple wind farm locations. 2 - Considerations for Modeling the Integration of Wind Power in the Electric Power Sector Kelly Eurek, National Renewable Energy Laboratory, Golden, CO, United States, Kelly.Eurek@nrel.gov Because wind power is a rapidly growing component of the electricity system, robust representations of wind technologies should be included in capacity- expansion models. This is a challenge because modeling the electricity system—and, in particular, modeling wind integration within that system—is a complex endeavor. This presentation highlights the major challenges of incorporating wind technologies into capacity-expansion models and shows examples of how NREL’s Regional Energy Deployment System (ReEDS) model address those challenges. TD19

382