Informs Annual Meeting 2017

SC67

INFORMS Houston – 2017

SC64

2 - Experience Visual Analytics with Tableau Ali Kherani, Tableau Software, Seattle, WA, United States, kherani@tableausoftware.com Get ready to learn how to inform and inspire others when presenting data with Tableau’s business intelligence platform. Join us for this tutorial to discover how to turn data into actionable insights and find out why Forbes recently ranked Tableau as the technical skill with the third biggest rise in demand. We’ll provide an overview of Tableau and demo the key features and functionality of our visual analytics platform. This overview will excite you to explore your data in a different way!

370E Data Analytics & Modeling in Medical Imaging Analysis & Decision Making Sponsored: Data Mining Sponsored Session Chair: Shouyi Wang, University of Texas at Arlington, Arlington, TX, 76013, United States, shouyiw@uta.edu 1 - Supervised Discriminative EEG Brain Source Imaging with GraphRegularization Feng Liu, University of Texas at Arlington, 1020 W. Abram Street, Apt 172, Arlington, TX, 76013, United States, feng.liu@mavs.uta.edu The advances in computing and advanced data analytics techniques have led to powerful and promising data-driven tools to solve complex healthcare and medical decision making problems based on increasingly large amounts of data. This session provides a collection of talks in the field to present several most recent research studies that develop data-driven approaches and machine learning techniques for comprehensive decision modeling in medical imaging analysis and decision making. 2 - Discriminative Spectral Pattern Analysis for Prostate Cancer Detection using Light Reflectance Spectroscopy Rahil Hosseini, University of Texas-Arlington, Arlington, TX, United States, rahilsadat.hosseini@mavs.uta.edu This talk presents an effective classification procedure that detects the positive margin in prostate cancer by application of data mining techniques. We developed a data-driven method to process and extract features from the Light Reflectance Spectroscopy (LRS) spectrum data, and developed support vector machine and ensemble trees approach for binary and multi-class classification for prostate cancer detection. 3 - A Supervised Learning Model for Predicting Tibia Soft Tissue Insertions using Multi-response Support Vector Regression Kin Ming Puk, University of Texas at Arlington, 711 Linda Vista Ave, Apt 112A, Arlington, TX, 76013, United States, pukkinming@gmail.com It has been an ongoing research to precisely identify the location of knee injury without affecting adjacent tissues during the surgical operation, subject to difficulties in locating ligament and related soft tissues and high inter-person morphological variability between patients. This work presents a new prediction framework which is capable of achieving personalized identification of cruciate ligament and optimal position of soft tissue insertions. It is made possible because tibia outline can be accurately and reliably measured from computed tomography imaging and features can thus be extracted for model building with L2,1- regularized multi-response support vector regression. 4 - A Network Modeling and Prediction Framework for Better Epilepsy Surgical Decisions Shouyi Wang, University of Texas-Arlington, 2029 Randy Snow Road, # 1310, Arlington, TX, 76011, United States, shouyiw@uta.edu Resective surgery is currently the most effective overall treatment for patients with epilepsy. However, there is still considerable risk of surgical failures for lacking of priori knowledge of seizure spread network. This talk will present a novel network modeling framework using cortico-cortical evoked potentials (CCEP) to understand network mechanisms of epileptogenesis and render clinicians better epilepsy surgical decisions in the future. 370F 1:30 - 2:15 Arundo/ 2:15 - 3:00 Tableau Invited: Vendor Tutorial Invited Session 1 - IIOT and Data-Driven Modeling in Asset Heavy Industry Tailai Wen, Arundo, Houston, TX, United States, tailai.wen@arundo.com, Roy Keyes The traditional industries start moving towards big data in recent years. While most acknowledge it may bring efficiency improvement and cost reduction to those asset-heavy sectors, to deploy data modeling infrastructure in an enterprise to industrial scale still faces challenges that were not seen in align and less complex modern industries before. In this talk, we will introduce the opportunities and obstacles in this field using real examples that Arundo works on, and demonstrate our solutions to these challenges. SC65

SC66

371A Autonomous Vehicles Sponsored: Transportation Science & Logistics Sponsored Session

Chair: Yiwen Xu, North Dakota State University, P.O. Box 6050, Industrial & Manufacturing Engineering, Fargo, ND, 58108-6050, United States, yiwen.xu6@gmail.com 1 - Recasting Interseciton Automation as a CAV Scheduling Problem within the Mixed Traffic Environment Pengfei (Taylor) Li, Assistant Professor, Mississippi State University, P.O. Box 9546, 501 Hardy Road, 235 Walker Hall, Mississippi State, MS, 39762, United States, pengfei.li@cee.msstate.edu, Xuesong Zhou Within the mixed traffic environment, connected-automated vehicles (CAV) are both part of traffic flow and green light request senders. We propose a new vehicle-centered traffic representation and a phase-time-traffic (PTR) hypernetwork model to optimize intersection automation policy (IAP). The new vehicle-centered traffic representation can reflect new features of CAVs in traffic flows. The PTR hypernetwork is to model the integration of mixed traffic with traffic signal operations. We also develop exact algorithms to reduce search space and search the exact optimal solution based on branch-and-bound framework. Algorithms are further enhanced with parallel computing technique. 2 - UAV Delivery Planning Considering Mission Failure Probability Yiwen Xu, Assistant Professor, North Dakota State University, PO Box 6050, Industrial & Manufacturing Engineering, Fargo, ND, 58108-6050, United States, yiwen.xu@ndsu.edu Commercialization of Unmanned aerial vehicles (UAVs) is more and more popular. As an emerging application, UAV delivery has attracted considerable attentions. Most previous research on the UAV delivery planning, however, did not consider the mission failure probability and the corresponding hazards. In this talk, we consider such factors and formulate the models to mixed integer programming (MIP) problems. The objective function is set to minimize the expected cost (including expected hazard costs) under given total mission time for all deliveries. We consider both deterministic demands and distributed demands. The models are solved by MIP solvers and numerical examples are included. 371B Machine Learning for Manufacturing Informatics Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Hao Yan, yanhao@gatech.edu Co-Chair: Jianjun Shi, Georgia Institute of Technology, Atlanta, GA, 30332-0205, United States, jianjun.shi@isye.gatech.edu 1 - Machine Learning via Nonconvex Model Based Optimization Tuo Zhao, Georgia Tech, 755 Ferst Dr. NW, Atlanta, GA, 30332, United States, tourzhao@gatech.edu, Han Liu, Tong Zhang Nonconvex optimization arises in many machine learning problems. Although classical optimization theory has shown that nonconvex optimization is computationally intractable, practitioners have proposed numerous heuristic algorithms, which achieve outstanding performance in real-world applications. To bridge this gap between practice and theory, we propose a new generation of model-based optimization algorithms and theory, which incorporate the statistical thinking into modern optimization. We exploit hidden geometric structures behind nonconvex optimization problems, and obtain global optima with desired statistics properties in polynomial time with high probability. SC67

91