Informs Annual Meeting 2017

TB35

INFORMS Houston – 2017

TB33

routing problem (MILIRP). We examine the unique structure of the MILIRP by considering a small problem instance and prove that the optimal policy is monotone when a sufficient penalty is applied. Moreover we show how the unique structure can be exploited by utilizing monotone approximate dynamic programming algorithms. 2 - Analysis of Personnel Policy via Discrete Event Simulation Lee Evans, University of Louisville, lee.evans@louisville.edu An effective performance appraisal system is critical in identifying officers with the knowledge, skills, and abilities to lead the future force. This study uses discrete event simulation to analyze the effect of system structure, system dynamics, and human behavior on the performance of the United States Army’s officer evaluation system. We demonstrate the efficacy of simulation optimization throughout the model development process, from model validation to model implementation. 3 - Using Markov Decision Processes with Heterogeneous Queueing Systems to Examine Military MEDEVAC Dispatching Policies Phillip R.Jenkins, Air Force Institute of Technology, WPAFB, OH, 45433, United States, phillip.jenkins@afit.edu, Matthew J.D. Robbins, Brian J. Lunday The goal of this research is to identify how to dispatch military medical evacuation (MEDEVAC) units to optimize system performance. A Markov decision process (MDP) model is developed to examine the MEDEVAC dispatching problem. A notional planning scenario is utilized to investigate differences between the optimal policy and three practitioner-friendly policies. Results indicate that dispatching the closest available unit is not always optimal. Moreover, an analysis of solution approaches for the MEDEVAC dispatching problem reveals that the policy iteration algorithm substantially outperforms the linear programming algorithms executed by CPLEX 12.6 in regards to computational effort. 351E Supply Chain Management, Green Contributed Session Chair: Samar K Mukhopadhyay, GSB SKK University, 53 Myeongnyung-dong 3-ga, Jongno-gu, Seoul, 110-745, Korezca, Republic of, samar@skku.edu 1 - Severity Assessment of Food Security Impediments in Food Supply Chain Rachita Gupta, Research Scholar, Indian Institute of Technology Delhi, New Delhi, 110016, India, rachitagupta1987@gmail.com, Ravi Shankar One of the major challenges worldwide is food security. The study proposes integrated Evidential Reasoning Algorithm to prioritize various food security impediments during logistics and distribution phases of Food Supply Chain. This would enable regulatory authorities to identify the most severe impediment and ensure affordable food security to the masses. 3 - Selection of Online Pickup Facility Networks using Fuzzy AHP Choonki Min, YonginSongdam College, Yoingin, Korea, zcRepublic of, ckmin@ysc.ac.kr, Byeong-Yun Chang Online pickup has become one of the most popular omnichannel services. In this study, we categorized the types of online pickup service facilities and analyzed the characteristics of each facility type. Using the fuzzy AHP method, the relative importance of online pickup service selection factors for each type of facility was derived from related experts. 4 - Managing Used Products: Who Should Refurbish? Narendra Singh, Indian School of Business, Knowledge City, TB35 We study a supply chain where a manufacturer sells new products through a retailer and decides whether to refurbish used products itself or to let the retailer refurbish them. We examine whether and when should a manufacturer allow the retailer to refurbish the used products. 5 - Brand Selection in a Supply Chain with Customer Returns Jing Chen, Professor, Dalhousie University, 6100 University Avenue, Halifax, NS, B3 H.4R2, Canada, JChen@dal.ca, Bintong Chen A retailer can sell a retailer can sell either or both of two brands, a well-known brand and a new brand, in a market supplied by two manufacturers in a manufacturer Stackelberg supply chain. We identify the conditions under which the retailer should choose one or both of the two manufacturers. We also show that an MBG enhances the profit of the manufacturer with low satisfaction rate, resulting in an increase in both the wholesale price and demand, but it has an opposite impact on the manufacturer with high satisfaction rate. In addition, an MBG enhances the retailer’s profit and expands the overall market. Sector 81, SAS.Nagar, Mohali, 140 306, India, narendra_singh@isb.edu, Ahmed Timoumi

351C Aviation Forecasting Sponsored: Aviation Applications Sponsored Session

Chair: Reed Harder, Dartmouth, 14 Engineering Drive, Hanover, NH, 03755, United States, reed.haseltine.harder.TH@dartmouth.edu 1 - Airline Entry and Exit Predictive Modeling: A Structural Econometric Approach Versus Machine Learning Chia-Mei (Mei) Liu, Federal Aviation Administration, Washington, DC, United States, Chia-Mei.Liu@faa.gov, Peng Wei, Xufang Zheng, Lei Kang Airline passenger route choice forecast modeling has not been a widely explored area in the airline forecasting community even though it is studied quite often in the airline empirical papers. This paper takes the initiative to build a model to forecast passenger choice between direct and connecting flights. Individual itineraries from 2000 to 2016 are aggregated to the airport-pair level to form our target variable, the percent of passengers flying direct. We compare forecast performance between statistical modeling and machine learning. Our preliminary findings show great promise with machine leaning techniques. 2 - Predicting Performance of Traffic Management Initiatives Alexander Estes, University of Maryland-College Park, College Park, MD, United States, aestes@math.umd.edu, Michael O. Ball, David J.Lovell The Federal Aviation Administration uses traffic management initiatives to manage excess demand for resources in the national airspace system. We present methods for estimating the performance that a given traffic management initiative will produce under a given set of weather and traffic conditions. These methods combine a weighting scheme from Geographic Weighted Regression with forest-based regression methods. Estimates of the conditional density can be produced along with standard regression estimates. 3 - Airline Entry and Exiting Predictive Modeling: A Structural Econometric Approach Kang Hua Cao, Hong Kong Baptist University, WLB 517, 34 Renfrew Road, Kowloon Tong, Hong Kong, kanghuacao@hkbu.edu.hk, Chia-Mei Liu, Vikrant Vaze Airlines’ decisions on entering or exiting a market are interrelated with other airlines’ choices of entry and exit. Ignoring such strategic effects in forecasting can lead bias results. In this paper, we develop a discrete game and estimate the structural parameters of the payoff function. Thus, we recover valuable information about how airline makes entry/exit decisions and predict its future outcomes. We apply the structural econometric approach on a rich airline dataset, which features variables signaling the health of the economy such as employment and housing starts, network variables such as the connectivity at the airports, and competition variables such as Herfindahl-Hirschman Index. 4 - Combining Empirical and Game Theoretic Approaches to Airline Frequency Decision Modeling Reed Harder, Dartmouth, 14 Engineering Drive, Hanover, NH, 03755, United States, reed.haseltine.harder.TH@dartmouth.edu, Vikrant Vaze Forecasts of air traffic are leveraged by the Federal Aviation Administration in planning and policy making. In practice, airlines allocate capacity across their networks in the face of competition from other airlines on service frequency and price. We present approaches to combining empirical data on airline operations and game theoretic models of airline competition for the purpose of modeling and forecasting airline frequency decisions, and show the potential of behavioral models of airline decision making to augment air traffic forecasts. 351D Military Decision Making Under Uncertainty Sponsored: Military Applications Sponsored Session Chair: Brian J. Lunday, Air Force Institute of Technology, Air Force Institute of Technology, WPAFB, OH, 45433, United States, brian.lunday@afit.edu 1 - A Military Variant of the Stochastic Inventory Routing Problem Ethan Salgado, Air Force Institute of Technology, Dayton, OH, United States, ethan.salgado@afit.edu, Matthew J.D. Robbins, Darryl Ahner Military commanders resupply forward operating bases from a central location within the area of operations mainly via convoys. Commanders must decide when and how much inventory to distribute throughout the area of operations while minimizing soldier risk; we call this problem class the military inventory TB34

315