Informs Annual Meeting 2017

SC01

INFORMS Houston – 2017

Sunday, 1:30 - 3:00PM

2 - Risky Choice Following Near Miss Events in Sequential Tasks under Ambiguity Florian Mathis Federspiel, Assistant Professor, INCAE Business School, Alajuela, Costa Rica, florian.federspiel@incae.edu, Matthias Seifert, Robin Dillon-Merrill Studies have shown that near miss events can lead to inconsistent risk perceptions. Yet near misses are often clouded in ambiguity, allowing for hubris and misattribution of what caused success or prevented failure. We provide an analytical model of near misses and investigate the experience of such an event on risk taking in a real options task. Over two experiments, we show that increases in risk taking following a near miss occur mainly under ambiguity. We further find that this effect depends on the decision maker’s prior expectation. Only those with an expectation of failure fall prey to the near miss bias. 3 - Group Decision Making with Multiple Conflicting Performance Targets Matthias Seifert, IE Business School, Maria de Molina 12, 5, Madrid, 28006, Spain, matthias.seifert@ie.edu, Enrico Diecidue, Anshdha Ria In many multi-stakeholder decision contexts, it is of crucial importance to aggregate preferences in a way that rewards the diversity of targets met. Extending Tsetlin & Winkler’s (2007) work on multiattribute performance targets, we study the properties of a novel group-weighting rule, which draws on the dis- /similarity of preferences held by individuals and assigns higher values to those alternatives that maximize the diversity of targets achieved. We then benchmark the performance of this rule against classic approaches based on equal weighting (e.g. average winner, majority voting). 4 - Dynamic Pricing under Habit Formation and Satiation Wen Chen, Providence College, United States, wchen@providence.edu, Ying He we study the dynamic pricing problem when consumer forms habit and/or satiation from their past consumptions in this paper. Our analysis is based on utility model under habit formation and satiation that are proposed and axiomatized in the decision theory literature recently, which provides a solid behavioral foundation for our model. Based on this utility model, we derive an inter-temporal dependent demand function to model how demand in each period depends on the prices in the current period and previous periods through habit formation and satiation. Our analysis shows that “U-shaped”, or “Inverse-U- shaped” policy may become optimal. 310C Easy Affine Markov Decision Processes: Properties and Applications Invited: Tutorial Invited Session Chair: Jiming Peng, University of Houston, Houston, TX, 6, United States, jopeng@Central.uh.edu Co-Chair: Rajan Batta, University at Buffalo (SUNY), 410 Bell Hall, University at Buffalo (SUNY), Buffalo, NY, 14260, United States, batta@buffalo.edu 1 - Easy Affine Markov Decision Processes: Properties and Applications Matthew J.Sobel, Case Western Reserve University - Retired, Department of Operations, 10900 Euclid Avenue, Cleveland, OH, 44106-7235, United States, matthew.sobel@case.edu, Jie Ning This tutorial introduces a class of decomposable affine Markov decision processes (MDPs) that have continuous multi-dimensional endogenous states and actions, and an exogenous state that follows an exogenous Markov chain. We show that, unlike most MDPs with continuous state and actions, decomposable affine MDPs are free of the curse of dimensionality and can be solved easily and exactly. A decomposable affine MDP has a value function that is affine in the endogenous state, and has an extremal optimal policy. We illustrate the potential applicability of decomposable affine MDPs using examples of fishery management and dynamic capacity portfolio management. SC03

SC01

310A Decision Analysis Practice Award Sponsored: Decision Analysis Sponsored Session Chair: Gregory L. Hamm, Stratelytics, LLC, Alameda, CA, 94501, United States, ghamm@strts.com 1 - A Methodology for Optimized Oil Well Location During Operational Drilling in Presence of Geological Uncertainty D. Echeverría Ciaurri, IBM.Researcch, Yorktown Heights, NY, United States, decheve@us.ibm.com, V.M. Babin, A.M. Vashevnik, O.S. Ushmaev, A.P. Gruzdev, A.V. Pozdneev, D. Echeverría Ciaurri We propose a novel methodology for the construction of an oil field development strategy that leverages decision trees and takes into account geological uncertainties of a reservoir. The approach is designed to dynamically decide on the drilling location of the next wells and on the need for appraisal wells. The strategy can be delivered to the drilling operator in the form of a pre-computed manual. The methodology was applied to a new oil field and 15% improvement in expected net present value was obtained with respect to an optimized development strategy with predetermined wells placement. 2 - How to Improve Educational Programs for Underprivileged Children? the Impacts of Value-focused Decision Analysis Valentina Ferretti, London School of Economics and Political This project develops a value-focused Decision Analysis intervention designed and deployed to support the Csányi Foundation in selecting underprivileged children to enter the Foundation’s Educational program. To minimize dropout rates, the Foundation expressed the need to improve the selection process, by better identifying the selection criteria that will allow the Foundation to meet its objectives and by fostering consensus among the board members. To achieve these targets, we combined a use of Multi Attribute Value Theory and Value Focused Thinking within a facilitated mode of engagement. 3 - Decision Analytics for Reducing Surgical-site Infections: Application to Coronary Artery Bypass Graft Patients Eva Lee, Georgia Tech, Industrial & Systems Engineering, Ctr for Operations Research in Medicine, Atlanta, GA, 30332-0205, United States, evakylee@isye.gatech.edu This study focuses on reducing SSI for coronary artery bypass graft(CABG) patients at a large hospital. A system-approach - using decision trees, machine learning, simulation and multi-objective optimization - was employed to account for the interdependency of preoperative, intraoperative and postoperative processes. A number of intervention and process changes were initiated. The site zero percentage within six months and sustained that rate for over 24 months. 310B Behavioral Decision Analysis Sponsored: Decision Analysis Sponsored Session Chair: Matthias Seifert, IE Business School, IE Business School, Madrid, 28006, Spain, matthias.seifert@ie.edu 1 - The Effectiveness of Trimmed Opinion Pools in Time Series Forecasting Involving Structural Breaks We introduce simple trimming approaches to aggregate judgmental time series forecasts involving structural breaks. While the extant literature explores trimming rules for aggregating forecasts in stable environments, we focus on relatively unstable environments characterized by fundamental regime shifts. In an empirical study, we find that forecasters are relatively under confident and sensitive to regime shifts and fluctuations in time series, making static trimming approaches less applicable. We further propose asymmetric trimming approaches to aggregate opinion pools under such unstable environments. SC02 Shijith Kumar Payyadakkath Meethale, PhD Candidate, IE Business School, Madrid, Spain, shijith.pm@Student.ie.edu, Matthias Seifert, Yun Shin Lee Science, Houghton Street| London | WC2A 2AE, London, United Kingdom, V.Ferretti@lse.ac.uk, Gabriella Csányi

66