2016 INFORMS Annual Meeting Program

WE38

INFORMS Nashville – 2016

WE39 207A-MCC Applied Probability and Economics II Sponsored: Applied Probability Sponsored Session Chair: Krishnamurthy Iyer, Cornell University, 225 Rhodes Hall, Ithaca, NY, 14853, United States, kriyer@cornell.edu 1 - The Magician’s Shuffle: Re-using Lottery Numbers For School Seat Redistribution Irene Yuan Lo, Columbia University, New York, NY, United States, iyl2104@columbia.edu, Itai Feigenbaum, Yashodhan Kanoria, Jay Sethuraman We consider a dynamic model of the school choice problem, where students are given an initial allocation, learn the value of their outside option, and are then given a final allocation based on their updated preferences. The goal is to obtain an efficient final allocation that is individually rational with respect to the initial allocation and minimizes student movement. We propose a family of mechanisms that are fair, efficient, two-round strategy-proof and individually rational. We show that when a natural ‘order’ condition is satisfied, all these mechanisms produce equivalent final allocations, and that the mechanism that reverses Ehsan Valavi, Columbia University, New York, NY, 30332-0205, United States, valavi19@gsb.columbia.edu, Costis Maglaras We study some promotional pricing heuristics in settings where consumers get informed about a product through a product reviews via a social learning mechanism. 3 - The Strange Case Of Privacy In Equilibrium Models Rachel Cummings, California Institute of Technology, 1200 E California Blvd, MC 305-16, Pasadena, CA, 91125, United States, rachelc@caltech.edu We study how privacy technologies affect user and advertiser behavior in a simple economic model of targeted advertising. In our model, a consumer first decides whether or not to buy a good, and then an advertiser chooses an ad to show her. The advertiser would like to use information about the consumer’s purchase decision to target the ad that he shows, but he is given only a differentially private signal about the consumer’s behavior. We study equilibrium behavior as a function of the privacy level and show that this behavior can be highly counter- intuitive. The effect of adding privacy in equilibrium can be completely different from what we would expect if we ignored equilibrium incentives. 4 - Delay-predictability Tradeoffs In Reaching A Secret Goal Kuang Xu, Stanford University, Stanford, CA, United States, kuangxu@stanford.edu, John N. Tsitsiklis We formulate a model of dynamic decision-making to study an agent’s predictability as she attempts to reach a final goal through a sequence of intermediate actions, while watched by an adversary who tries to predict the goal before it is reached. We are motivated by the increasing ubiquity of large-scale data collection infrastructures capable of predicting an agent’s intentions and future actions, in juxtaposition with an agent’s desire for privacy. We show the predictability of the agent’s goal can be made inversely proportional to the time she spends reaching it, and that this is the best possible. This characterization does not depend on the structure of the agent’s state space beyond the diameter. student lotteries between rounds minimizes student movement. 2 - Promotional Campaigns Under Social Learning Queues with Correlations Sponsored: Applied Probability Sponsored Session Chair: Ohad Perry, Northwestern University - Evanston, Evanston, IL, United States, ohad.perry@northwestern.edu 1 - The Impact Of Delays On Service Times In The Intensive Care Unit Carri Chan, Columbia Business School, cwchan@columbia.edu Most queueing models used to model healthcare delivery ignore the effects of delay experienced by patients awaiting care. We empirically verify that delays in time-to-treatment can increase a patient’s service requirement. We then propose a queueing model which incorporates these measured delay effects and approximation the expected work in the system when the service time of a job is adversely impacted by the delay experienced by that job. Our approximation demonstrates that work grows much faster than the traditional 1/(1 − ) relationship seen in most queueing systems. As such, ignoring this effect of delay could have dire operational consequences. WE40 207B-MCC

2 - Modeling Medical Overpayments Using Truncated Distributions Babak Zafari, Babson College, Babson, MA, United States, zafari.babak@gmail.com In this work, we explore some new methods used in overpayment extrapolations and compare their performance to existing models. 3 - Procurement Models For Clinical Supplies: Indian Context Bhavin J Shah, Associate Professor, Indian Institute of Management, Indore, Faculty Office # C-206, First Floor,, Prabandh Shikhar, Rau-Pithampur Road,, Indore - Madhya Pradesh, 453556, India, bhavinj@iimidr.ac.in, Hasmukh Gajjar This paper seeks to understand and explore applications of procurement models followed in sourcing clinical supplies to reduce cost of healthcare in Indian hospitals. It aims at improving efficiency of healthcare delivery without sacrificing service levels and explore various alternatives such as forming cross-functional collaborative teams comprising clinicians and sourcing experts for operational improvements. 4 - A Multiple Criteria Decision Tree Algorithm For Selecting Breast Cancer Treatment Mostafa Hasan, Research Assistant, Wichita State University, 1629 N Fairmount St, Wichita, KS, 67208, United States, mhasann16@yahoo.com, Esra Buyuktahtakin, Elshami Elamin According to the American Cancer Society, 246,660 new cases will be diagnosed with invasive breast cancer and approximately 40,450 women will die in the United States in 2016. To deal with the complexity, a decision support system is proposed combining MCDM techniques and decision trees. We then propose a detailed algorithm which will evaluate each factor and condition of the breast cancer patients in order to determine the best treatment alternatives. WE38 206A-MCC General Session III Contributed Session Chair: Seyedali Mirzapour, PhD Student, Wichita State University, 1845 Fairmount St., Wichita, KS, 67260, United States, mirzapour.ie@gmail.com 1 - Modeling And Performance Evaluation Of Bernoulli Transfer Lines With Batch Processors Feiyi Yan, Northwestern Polytechnical University, 127 West Youyi Road, Xi’an, China, pacpos.fyyan@gmail.com, Jun-Qiang Wang This paper focuses on analytical methods for performance evaluation of transfer lines with Bernoulli unreliable batch processors and finite buffers. Each machine has a limited capacity to process a batch of parts simultaneously. Three different analytical models of production lines with two batch processors are established. The modulo n congruence class theory is introduced to depict the system state of the Markov chain and to prove the ergodicity condition. An aggregation method is proposed to analyze a general Bernoulli transfer lines. Numerical experiments are conducted to verify the accuracy of the proposed methods. The impact of machine capacity on system performance is analyzed. 2 - A New Relaxation Method For Mathematical Program With Complementarity Constraint Tangi Migot, IRMAR-INSA, Rennes, France, tangi.migot@gmail.com, Jean-Pierre Dussault, Mounir Haddou p { margin-bottom: 0.25cm; line-height: 120%; } Recent progress on optimality conditions for MPCC allows to build efficient relaxation methods starting from Kadrani et al. in 2009. We will present an overview on these methods and discuss both the properties of the sequence of non-linear program (NLP) generated by these algorithms and the weakest conditions needed to ensure convergence of the methods. We will also present a new method with improved properties on the sequence of NLP, which provides a certificate when the method converges to undesirable points. We run a numerical comparison of these methods on a large number of test problems. 3 - Leaf Trajectory Optimization For Dynamic Delivery Of Intensity-modulated Radiotherapy Plans Seyedali Mirzapour, PhD Student, Wichita State University, 1845 Fairmount St., Wichita, KS, 67260, United States, mirzapour.ie@gmail.com, Ehsan Salari In intensity-modulated radiotherapy, a multi-leaf collimator (MLC) consisting of rows of paired leaves, is used to dynamically modulate the shape and intensity of radiation beams. Traditionally, unidirectional leaf sweeping schemes have been considered for dynamic beam modulation. In this research, we investigate the potential gain in plan quality and efficiency obtained from allowing for a free movement of MLC leaves.

496