2016 INFORMS Annual Meeting Program

WB44

INFORMS Nashville – 2016

WB45 209A-MCC Simulation for Performance of Non-Stationary Queues Sponsored: Simulation Sponsored Session Chair: John Shortle, George Mason University, Fairfax, VA, United States, jshortle@gmu.edu 1 - A Performance Algorithm For Periodic Queues

3 - Operations Decisions With Target Based Incentives Andrea Hupman Cadenbach, University of Missouri - St. Louis, cadenbach@umsl.edu Supply chains represent complex systems in which numerous performance measures are used and in which responses to incentive structures can change system performance. One type of incentive is a fixed target in which performance above a threshold is rewarded. The literature shows that high targets tend to induce risk taking while low targets tend to induce the selection of safer decision alternatives, but this talk examines the effects of setting targets on different types of performance measures in addition to the magnitude of the target. 4 - Team Incentives In Multi-level Principal Agent Problems Aditya Umesh Kulkarni, PhD Candidate, Virginia Tech, 11700 Cardinal Court, Apt E, Blacksburg, VA, 24060, United States, aditya88@vt.edu, Christian Wernz Existing literature on relational contracts typically does not account for the interdependency of stochastic performance outcomes across multiple levels of superior-subordinate interactions in organizations. We use multiscale decision theory to derive relational, that is, trust-based, contracts for teams while accounting for the aforementioned influences. We derive the optimal job design for employees, which is a function of rewards and performance influence. Our results extend agency theory by accounting for cascading and team-based incentives in organizations. WB44 208B-MCC Behavioral Decision Analysis Sponsored: Decision Analysis Sponsored Session Chair: Matthias Seifert, IE Business School, Maria de Molina 12, 5, Madrid, Spain, matthias.seifert@ie.edu 1 - Exploring The Consistency Of Higher-order Risk Preferences Timo Heinrich, IN-EAST School, Universität Duisburg-Essen, Duisburg, Germany, timo.heinrich@ibes.uni-due.de, Alexander Haering, Thomas Mayrhofer We measure higher-order risk preferences and explore their consistency across orders. We analyze the role of (i) country differences between China, Germany, and the US, (ii) differences in stake size, and (iii) differences through displaying reduced rather than compound lotteries. We replicate the finding of mixed risk- averse and mixed risk-loving behavior by Deck and Schlesinger (2014) in the US and identify a similar pattern in Germany and in China. Moreover, we observe an increase in risk aversion when stakes are increased tenfold. Finally, in reduced lotteries there is only weak evidence for prudence and no evidence for temperance. 2 - Regime Shift Detection In The Domain Of Losses Matthias Seifert, Associate Professor of Decision Sciences, IE Business School, Madrid, 28006, Spain, matthias.seifert@ie.edu, Sara Farooqi We extend Massey and Wu’s (2005) work on regime shift detection by studying the influence of system neglect on trading behavior in experimental markets. We show in a series of laboratory experiments how patterns of probabilistic over- and underreaction translate into investment decisions under risk and use Prospect Theory to explain systematic differences between buyers/sellers as well as gains/loss domains. 4 - Risky Choice Following Near Miss Events In Sequential Tasks Under Ambiguity Florian Federspiel, IE Business School, Madrid, Spain, fmfederspiel@faculty.ie.edu, Matthias Seifert 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. 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 - The Reasonability Of Behavioral Assumptions Made In Complex Systems Models Allison C Reilly, University of Michigan, Ann Arbor, MI, United States, allison.reilly@gmail.com, Seth Guikema Improved capture of human behavior in complex systems modeling has significantly advanced in recent decades and proffers a more cohesive approach to understanding how these systems may operate, fail, or evolve. The behavioral assumptions have significant implications on the insights derived from these models, though these implications are rarely explored. In this work, we address frequently confused decision science terminology used in systems models - from rationality to strategy - the reasonability of the behavioral assumptions, and their implications via case studies.

Ni Ma, Columbia University, nm2692@columbia.edu, Ward Whitt An efficient algorithm is developed to calculate the steady-state distribution of virtual waiting time in a general Gt/G/1 queue with a periodic arrival-rate function. We first approximate the Gt/G/1 model by an associated GIt/GI/1 model based on a recent heavy-traffic functional central limit theorem, and then compute the exact tail probabilities of the virtual wait by exploiting a modification of the classic exponential change of measure. That algorithm is then applied to compute related performance measures, such as the mean and variance of the virtual wait. 2 - Fourier Trajectory Analysis For Identifying System Congestion Russell R Barton, Pennsylvania State University, 210 Business Building, University Park, PA, 16802, United States, rbarton@psu.edu, Xinyi Wu We examine the use of the Fourier transform to discriminate dynamic behavior differences between congested and uncongested systems. Simulation continuous time statistic ‘trajectories’ are converted to time series for Fourier analysis. We use this knowledge to explore statistical process control methods to monitor nonstationary systems for transition from uncongested to congested state and vice versa. 3 - Staffing To Stabilize Blocking In Loss Models With Time-varying Arrival Rates Jingtong Zhao, Columbia University, jz2477@columbia.edu, Ward Whitt It is not possible to find a time-varying staffing policy to stabilize blocking probabilities in a multiserver loss model with a time-varying arrival rate to the same extent as in corresponding delay models, because the blocking probabilities necessarily change dramatically after each staffing change, but nevertheless a variant of the established modified-offered-load staffing algorithm performs well if we randomize appropriately. 4 - Effects Of Arrival Variability on Delays at Congested Airports John Shortle, George Mason University, jshortle@gmu.edu More precise spacing of flight arrivals into airports has the potential to increase capacity and reduce delays. However, even with precise spacing, delays can also result from the variability in the mean arrival rates throughout the day (i.e., the non-stationary nature of the arrival process) due to banking of flights at hub airports. This talk presents a queueing simulation of an airport and investigates the relative impact of reducing the uncertainty in the spacing of arrivals versus reducing the schedule variability. Under conditions of high utilization, reducing the arrival variability has limited impact on delays unless also accompanied by reduced schedule variability.

WB46 209B-MCC

Networks and Games in Operations Sponsored: Revenue Management & Pricing Sponsored Session

Chair: Kimon Drakopoulos, Marshall School of Business, 3670 Trousdale Pkwy, Los Angeles, CA, 90089, United States, kimondr@mit.edu 1 - Information Obfuscation In Strategic Experimentation Kimon Drakopoulos, Massachusetts Institute of Technology, kimondr@mit.edu In our work we model a continuous time strategic experimentation problem against an informed opponent (incumbent) who can take actions to obfuscate learning. We show that the unique (weak) perfect Bayesian equilibrium of this dynamic game takes the form of the ”delayed war of attrition” where over a range of beliefs of the experimenter, both players use mixed strategies — for the entrant, to determine when to stop experimenting and for the incumbent, to determine when to stop obfuscating. The uniqueness of the outcome gives rise to many questions such as market design in order to maximize fairness for the entrants, or consumer surplus.

410