Informs Annual Meeting 2017

SC68

INFORMS Houston – 2017

2 - Machine Learning for in SITU Quality Monitoring of Manufacturing Processes Bianca Maria Colosimo, Politecnico di Milano, Via La Masa, 1, Milan, I-20156, Italy, biancamaria.colosimo@polimi.it, Marco Grasso In-situ monitoring of advanced manufacturing processes requires monitoring signal and image data with time. The contribution describes how machine learning approaches can be usefully included in the SPC procedure to detect unnatural process conditions. 3 - Modeling the Distribution of Nonlinear Profile with Application for Process Monitoring and Denoising Hao Yan, 3525 Highgrove Way NE, Atlanta, GA, 30319, United States, yanhao@gatech.edu Unsupervised learning of nonlinear profiles has attracted increasing attention among researchers as well as practitioners because of its applications in process monitoring and diagnostics. However, existing unsupervised learning techniques for profile monitoring are typically based on linear models such as Principal Component Analysis, which fails to model the distribution of nonlinear profiles. We present a unified framework for nonlinear profile modeling framework based on variational inference, which a testing statistics can be naturally derived. We then demonstrate how this framework can be used for denoising and profile monitoring with various simulations and case studies. 371C QSR Student Interaction Session Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Chiwoo Park, Florida State University, Tallahassee, FL, 32310-6046, United States, cpark5@fsu.edu Co-Chair: Linkan Bian, Mississippi State University, MS, 39762, United States, bian@ise.msstate.edu 1 - QSR Student Interaction Linkan Bian, Mississippi State University, MS, 39762, United States, bian@ise.msstate.edu The Student Introduction and Interaction Session is designed for QSR student members to get acquainted with each other, promote themselves in front of our community, and learn from invited guests. In this session, each student will be given two minutes to deliver an elevator speech about his/her research interests and accomplishments; senior QSR members and guests will be invited to interaction with all attendees. A parallel Best Student Poster Competition will be hold at the same room. Students are encouraged to prepare poster presentations to have better exposure. All posters will enter the competition automatically. A winner will be selected by a panel of judges, announced at the QSR business meeting, and awarded a certificate. 371D Food-Energy-Water and Sustainability Sponsored: Energy, Natural Res & The Environment Environment & Sustainability Sponsored Session Chair: Urmila Diwekar, Vishwanathan Research Institute, urmila@vri-custom.org 1 - Stochastic mixed-integer Programming (MIP) to the Competition of Biofuel and Food Production Esra Buyuktahtakin, New Jersey Institute of Technology, Newark, NJ, United States, esratoy@njit.edu, Halil Cobuloglu We present a two-stage stochastic MIP model that maximizes the economic and environmental benefits of food and biofuel production under yield and price uncertainty. The first-stage variables define binary decisions for allocating various land types to food and energy crops, while the second-stage variables are operational decisions related to harvesting, budget allocation, and amounts of different yield types. Results indicate the significant benefit of using the stochastic yield-level information in an optimization model that is solved by a Benders’ decomposition algorithm. SC68 SC69

2 - Optimizing Spatiotemporal Sensors Placement for Nutrient Monitoring: An Algorithmic Framework Urmila Diwekar, Vishwamitra Research Institute, 2714 Crystal Way, Crystal Lake, IL, 60012-2224, United States, urmila@vri-custom.org, Rajib Mukherjee Nutrient monitoring is very important for the area of food-energy-water nexus. In this work, we have proposed a methodology to optimize a dynamic sensor network which can address the spatiotemporal aspect of nutrient movement in a water shed. This is a first paper in the series where an algorithmic and methodological framework for spatiotemporal sensor placement problem is proposed. This framework is based on a novel stochastic optimization algorithm called Better Optimization of Uncertain Systems (BONUS) and is very efficient. A small case study of the dynamic sensor placement problem is presented to illustrate the approach. 3 - Limits to Growth and Global Sustainability of Food-energy-water Nexus Urmila Diwekar, Vishwamitra Research Institute, Crystal lake, IL, United States, urmila@vri-custom.org, Heriberto Cabezas A mathematical representation of World dynamics has been constructed inspired by the Limits to Growth. It includes an ecosystem, human population, industry, energy generation, water and the food-energy-water nexus, macro-economy, and rudiments of a legal system with private property, taxation, and regulation. The model provides the conceptual machinery for “steering” the system towards sustainability using socio-economic policies. Optimal control theory is used to forecast policies for three versions of the model. Analysis of sustainability with various scenarios is discussed, and controllability and socio-economic policies to bring the system towards sustainability are described. 371E Big Data Contributed Session Chair: Anurag Agarwal, University of South Florida, Sarasota, FL, United States, agarwala@usf.edu 1 - Real Time Detection of Acute Hypotensive Episode in Intensive Care Unit to Reduce Mortality Rate Rupesh Agrawal, Research Assistant, Oklahoma State University, SC70 Acute Hypotensive Events (AHE) requires timely attention for intensive care unit (ICU) patients to keep mortality rate low. Timely detection and trend of AHE from “normal” state to “not-normal” state can help ICU medical staff for timely intervention. This study uses rule-based “Complex Event Process” (CEP) detection method to calculate real-time state (normal/not-normal) of “Arterial Blood Pressure” (ABP) signal in real-time. ABP signal streaming at the rate of 125 Hz is used for this study to calculate AHE per second vs per minute, allowing medical Sharmin Nahar Mithy, Doctoral Candidate, University of South Florida, 2208 Fitness Club Way, Apt 204, Apt: 204, Tampa, FL, 33612, United States, sharminmithy@mail.usf.edu, Grisselle Centeno More than one million people get affected by colorectal cancer yearly. The best estimation of prognosis in colorectal cancer is related to the anatomic extent of disease determined on pathologic examination of the resection specimen. Development of statistical models to help the clinicians who will estimate the probability of developing cancer is of great interest in today’s research. In this research a neuro-fuzzy approach will be considered in the areas of classification of microarray data analysis. By combining the advantages of both Neural Network and Fuzzy Inference System, Nuero-Fuzzy approaches will be able to achieve a David Perkins, Associate Professor of Business, Grand Canyon University, 717 East Country Plaza North, Gilbert, AZ, 85234, United States, david.perkins@cox.net The Data Mining Cost Model (DoCoMo) attempts to provide a robust parametric approach for estimating the cost of a data mining project. This presentation will provide a review of important literature associated with the DoCoMo, will suggest how the DoCoMo can be applied in today’s analytics organizations, and will recommend further areas of application and research. professionals between 10-30 mins to provide timely intervention. 2 - Microarray Data Analysis in the Prognosis of Cancer higher accuracy within a relatively shorter training time. 3 - Insights into the Data Mining Cost Model 700 N. Greenwood Ave., Tulsa, OK, 74106, United States, rupesh.agrawal@okstate.edu, Surya B. Ayyalasomayajula, Dursun Delen, Ramesh Sharda, Bruce Benjamin

92