Informs Annual Meeting 2017

POSTER COMPETITION

INFORMS Houston – 2017

57 - Signal Control Strategy to Improve Person Mobility and Air Quality Farnoush Khalighi, PhD Student, University of Massachusetts, Amherst, 99 wildflower dr, amherst, MA, 01002, United States, fkhalighi@engin.umass.edu, Eleni Christofa This study develops an analytical model to estimate emissions produced by cars and buses at signalized intersections. Then, a real-time signal control system to minimize person delay and vehicular emissions is developed. The objective of the optimization problem presents different weighted combinations of person delay and vehicular emissions. A Pareto frontier of optimal solutions is generated to demonstrate the trade-off between person delay and emission levels and provide policy makers with a tool to choose between different signal control systems based on traffic conditions and air pollutant levels. This system provides priority to buses due to the higher weight given to them in the model. 58 - The Case for Nearliers Jeffry N. Savitz, SavitzConsulting, LLC, 1500 Locust St. #2120, Philadelphia, PA, 19102, United States, jsavitz@savitzresearch.com Nearliers are data points within a random sample whose values are closer to the sample mean than the average data point. In an empirical study, 719 respondents’ average ratings were computed for 30 popular brands. The average ratings of the Nearliers was within an absolute percentage difference of only 2.5% with the variance of the Nearliers being 17.4% less than the random sample. Thus, accurate and reliable predictions of population means can be made using Nearliers with a sample size 17.4% less than a random sample at an annual cost savings to the U.S. survey research community of $1.7 billion dollars based on information from CASRO and ESOMAR. Chair: Jeffrey W. Herrmann, University of Maryland-College Park, Department of Mechanical Engineering, College Park, MD, 20742, United States, jwh2@umd.edu Co-Chair: Sergiy Butenko, Texas A&M University, 4037 Emerging Technologies Building, Mail Stop 3131, College Station, TX, 77843- 3131, United States, butenko@tamu.edu 1 - Maximizing the Probability of Reaching a Set of States While Avoiding a Forbidden Set Daniel Felipe Ávila, Universidad de los Andes, Bogotá, Colombia, df.avila353@gmail.com, Mauricio Junca Using average cost functions we present a solution to the problem of finding a control policy that maximize the probability of reaching a set A while avoiding a set B. We show discounted cost functions provide an approximation to the solution. Both approaches can be solved using a linear program, the discounted case happens to admit an easier linear program than the average. We present simulations for the control of a plane under stochastic wind. 2 - Airline and Passenger Incentive Optimization Models for Airport Congestion Mitigation John Park, Assistant Professor, North Carolina A&T.State University, 800 revolution mill drive, apt 232, greensboro, NC, 27405, United States, hpark1@ncat.edu Bilevel Optimization Extended to Anticipatory Stochastic to Accommodate the Uncertainty. Passenger accessing based on airport with check in and security scheduling multiple airports with incentives. 3 - Estimating an Inverse Mean Subspace Jiaying Weng, University of Kentucky, 300 Alumni Drive, Apt 179, Lexington, KY, 40503, United States, jiaying.weng@uky.edu, Yin Xiangrong Estimating an inverse regression space is specially important in sufficient dimension reduction, and it typically requires a tuning parameter such as a number of slices in slicing method or bandwidth selection in kernel estimation approach, which then increases difficulties for multivariate responses. In this paper, we use Fourier transform idea to avoid such difficulties while easily incorporates with multivariate responses. We further develop Fourier transform method to deal with a sparse issue, categorical predictor variables and large p, small n data. Asymptotic as well as the sparse eigen-decomposition method for dimensionality tests are obtained. Monday Poster Competition Exhibit Hall Monday Poster Competition Competition Poster Session

4 - A Semantic Text Mining Technique for Classification of Manufacturing Suppliers Based on Their Capabilities Ramin Sabbagh, Graduate Research Assistant, Texas State University, 601 University Dr., San Marcos, TX, 78666, United States, r_s343@txstate.edu This work proposes a novel framework for capability-based supplier classification that relies on the unstructured capability narratives available on the suppliers’ websites. Naïve Bayes is used as the text classifier technique. One of the innovative aspects of this work is incorporating a thesaurus that contains the informal vocabulary used in the contract manufacturing industry for advertising manufacturing capabilities. An Entity Extractor tool is developed for the generation of the vector model associated with each capability narrative. The proposed framework is validated experimentally through forming two capability classes (heavy & complex machining) based on real capability data. 5 - Data Enabled Planning of Transcatheter Aortic Valve Replacement Procedure using Logistic Regression Model and 3d-printing Geet Lahoti, Graduate Research Assistant, ISyE, Georgia Institute of Technology, 1038 Mcmillan St. NW, Atlanta, GA, 30318, United Paravalvular leak (PVL) is a defect around the valve replacement, through which blood can leak. Prior assessment of PVL is preferable as it helps in procedural planning of TAVR. Patient-specific features can help predict PVL. We develop a logistic regression model for predicting PVL using computed tomography heart valve data. Five significant features including annulus diameter, none calcium, patients’ basic information (e.g. age and valve size) and the ratio of annulus diameter and valve size are used as predictors. Considering 178 patients, the accuracy of the developed model for training data is found to be 75.16% and that for testing data is 69.22% after performing 10-fold cross validation. 6 - Flexcct: A Methodological Framework and Software for Ratings Analysis and Wisdom of the Crowd Applications Stephen L. France, Assistant Professor, Mississippi State University, 324 McCool Hall, 40 Old Main, Mississippi State, MS, 39762, States, glahoti6@gatech.edu, Zih Huei Wang, Kan Wang, Zhen Qian, Mani Vannan, Chuck Zhang, Ben Wang FlexCCT provides an integrated framework and tool-set for analyzing and aggregating ratings. It utilizes a likelihood based statistical model to create aggregate ratings weighted for rater competencies and rater biases. It has features for the analysis of multiple rating cultures and for consensus adjusted reliability. Multiple optimization algorithms are implemented, along with a range of identifiability options to restrict certain subsets of the model parameters. Bootstrapping and jackknifing based methods are implemented to give confidence intervals for parameters. A holdout validation method is developed to allow for the testing of solution reliability and stability. 7 - Reducing Discarded Organs and Improving Physician-patient Decision-making via Decision-support Tools and Systems Optimization Ethan Mark, Georgia Institute of Technology, Atlanta, GA, United States, emark3@gatech.edu, David Goldsman, Pinar Keskinocak, Hannah Smalley, Joel Sokol Organ supply is not sufficient to meet demand for transplant. Organs with increased risk for Hepatitis B, Hepatitis C, HIV, and infectious Encephalitis could help fill the gap. We developed decision-support tools which estimate infection risks for donor organs and personalized post-transplant and waitlist survival curves for potential transplant recipients. In some cases, a potentially infected organ provides higher survival probability versus remaining on the waitlist. The decision support tools encourage shared and informed decision making between physicians and patients and may decrease the number of discarded organs by quantifying the tradeoffs of using a potentially infected organ. 8 - Improving the Resilience of Service System by a Network Flow Approach Jing Gong, Jinan University, School of Public Administration, Guangzhou, China, tgongjing@jnu.edu.cn Recent catastrophic events precipitate a wide range of impacts so that public awareness continues to grow of resilience as an important means of characterizing the ability to deal with them. The best service systems aren’t just cost-effective. They also respond quickly, handle smoothly and recover promptly from disruptions. This paper proposes a linear piecewise recovery function and applies it to the decision-making processes of improving a service system’s resilience to disasters. A numeric example is developed to demonstrate the proposed model enables tradeoff between a cost minimum system and a resilient system. United States, sfrance@business.msstate.edu, Mahyar Sharif Vaghefi, William H. Batchelder

200