Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TE41

2 - Substitutability, M-natural-convexity and Their Applications Menglong Li, UIUC, Urbana, IL, United States, Xin Chen Substitutability, an important concept in economics and operations research, poses significant technical challenges. In this paper, we build fundamental properties of M-natural-convexity and its variant SSQM-natural-convexity, and apply them to analyze operations models with substitution: a multi-product multi-period stochastic inventory model and a portfolio contract model where a buyer reserves capacities in blocks from multiple competing suppliers. 3 - Stochastic Optimization with Decisions Truncated by Random Variables and its Applications in Operations Xiangyu Gao, The Chinese University of Hong Kong, Hong Kong, Xin Chen, Zhan Pang We study a class of stochastic optimization problems with decisions truncated by random variables. We develop a transformation technique to convert the original non-convex problems to equivalent convex ones. Our transformation allows us to prove the preservation of some desired structural properties, such as convexity, submodularity, and L-natural-convexity, under optimization operations, which are critical for identifying the structures of optimal policies and developing efficient algorithms. We demonstrate the applications of our approach to several important models in operations management. n TE39 North Bldg 226A Joint Session MSOM/Practice Curated: Empirical Studies in Service Operations Sponsored: Manufacturing & Service Oper Mgmt/Service Operations Sponsored Session Chair: Song-Hee Kim, University of Southern California, Los Angeles, CA, 90089, United States 1 - Labor Planning and Shift Scheduling in Retail Stores Using Customer Traffic Data Marcelo Olivares, Universidad de Chile, Beauchef 851, Santiago, Chile This work combines empirical analysis of point-of-sales data, customer traffic and employee staffing, with optimization methods to build a decision support tool that can be used by managers to plan labor allocation and working schedules, balancing gross margins, labor costs and employee satisfaction. An econometric model is developed to estimate the effect of labor on sales in an hourly basis, decomposing the effect of labor into conversion and ticket value. The empirical results are used in mathematical program that seeks to find the best feasible schedule to maximize store profitability, accounting for labor regulation restrictions and practical constraints. 2 - Sooner or Later? Learning from Delivery Speed Information Ruomeng Cui, Emory University, 1935 Ridgemont Lane, Decatur, GA, 30033, United States, Tianshu Sun Online retailers who sell physical products need to inform customers how fast an order can be delivered. The delivery speed is a crucial service metrics that online retailers can easily adjust on the website. On one hand, an aggressive (faster) delivery estimate could meet more demand and thus may increase current sales. On the other hand, it may also raise the likelihood of customers experiencing a loss in time due to a longer than expected wait, which might hurt future sales. Collaborating with an online store selling custom products, we leverage the exogenous shipping policy change and use a difference-in-difference approach to examine whether and how consumers respond to the delivery information. 3 - Spatially Differentiated Services: Bike-share, Ride-hailing, Etc. Karan Girotra, Cornell Tech, New York, NY, 10044, United States We present the the key common patterns from our findings from collaborations with bike-share operators in Paris and London, ride-hailing companies and a shuttle service operator. We find that there are consistent patterns of commuter behavior and preferences in all these systems and use these patterns to provide guidance on the design of such systems. n TE40 North Bldg 226B Systems Engineering Decision Analysis Sponsored: Decision Analysis Sponsored Session Chair: Christian Wernz, Virginia Commonwealth University - VCU, P.O. Box 980205, Richmond, VA, 23298, United States, Virginia Tech, Virginia Tech, Blacksburg, VA, 24060, United States

1 - Trade-off Analytics with Quantitative Set-based Design Eric Specking, University of Arkansas, 4207 Bell Engineering Center, Fayetteville, AR, 72701, United States, Gregory S. Parnell, Edward A. Pohl Traditional system design (Point-Based Design), involves identifying and comparing a small number of design solutions. Set-Based Design (SBD) is an alternative approach that evaluates many points in sets of design choices. Repeatable quantitative techniques to perform SBD trade-off analysis of systems do not exist. This presentation will provide insights into SBD quantification literature and demonstrate mathematical methods to perform SBD trade-off analytics. These SBD techniques are enabled by an integrated MBE framework for system trade-off analytics using models and simulations. 2 - Multi-objective Decision Analysis in Transportation System Planning Edward A. Pohl, Professor, University of Arkansas, Department of Industrial Engineering, 4207 Bell Engineering Center, Fayetteville, AR, 72701, United States, Bobby Cottam, Gregory S. Parnell Decision analysis is important for designing an optimal transportation system. With a large variety of stakeholders, different types of users, and unique a set of considerations for every project multiple objective decision analysis is a useful tool to understand the tradeoffs and values of different alternatives in transportation planning. Resilience is another important consideration in transportation planning, between accidents weather and special events there are numerous times when a system is not functioning at optimal capacity, considering which alternatives perform best under adverse conditions can allow for a consistently higher level of service for travelers. 3 - An Alternative to Heat-map Risk Matrices for Project Risk Prioritization Jamie Peter Monat, Professor of Practice, Worcester Polytechnic Institute, 100 Institute Road, Worcester, MA, 01609-2280, United States, Scott Doremus Heat map risk matrices are widely used to prioritize risks. They are widely used in projects and programs structured for product development, business, construction, R & D, medicine, and politics. However, they suffer from several weaknesses, notably subjectivity, symmetry, category prioritization reversal, and a failure to account for Risk Aversion. Because of this, they can lead to serious mis- prioritization and mismanagement of risks. We propose an alternative to the standard risk matrix that accounts for these deficiencies and provides a significantly better risk prioritization tool. Two field trials were conducted to demonstrate the benefits of the alternative method. 4 - Optimal Verification Strategies in Multi-firm Projects: A Multiscale Decision Theory Approach Aditya Umesh Kulkarni, Virginia Commonwealth University, 1008 East Clay Street, Box 980203, Richmond, VA, 23298-0203, United States, Alejandro Salado, Christian Wernz Incentivizing verification activities of sub-contractors in multi-firm projects is a significant challenge for the main contractor. We present a belief-based model of verification that utilizes the multiscale decision theory (MSDT) modeling approach to determine optimal incentives and verification strategies in a multi- firm design project. Exact solution algorithms that determine optimal incentive and verification strategies are then presented for multiple scenarios. Our work contributes to MSDT and systems engineering literature by studying the hitherto unexplored problem of optimal verification and incentive strategies in multi-firm scenarios. n TE41 North Bldg 226C Decision Analysis and Health Care Policy Sponsored: Decision Analysis Sponsored Session Chair: Manaf Zargoush 1 - Predicting Health and Economic Outcomes of Providing Subsidy for Contraception in British Columbia, Canada Saied Samiedaluie, University of Alberta, 3-40C Business Building, Alberta School of Business, Edmonton, AB, T6G 2R6, Canada, Steven Shechter, Wendy Norman, Stirling Bryan Almost half of pregnancies in Canada are unintended. Cost is among the major barriers to effective family planning methods and services. Extending government subsidy for contraception has proven to be a cost-effective intervention in many countries. However, in Canada there is a paucity of evidence to support such programs. In this research, we developed a simulation model that can predict the related health and economic cost and benefit streams likely to result from any policy change in this area, such as provision of universal subsidy for contraception.

393