Informs Annual Meeting 2017

MA44

INFORMS Houston – 2017

4 - Uncertainty Analysis of Species Distribution Models Xi Chen, University of Texas at Austin, 3500 Greystone Drive, Apt 271, Austin, TX, 78731, United States, carol.chen@utexas.edu Species distribution models (SDM) are commonly used to predict the geographic distributions of animals or plants species. However, most of SDM methodology produces point estimates. We present two strategies for quantifying the uncertainty of maximum entropy SDM methods: a bootstrap method, and an analytic method. We describe both methods, and compare and contrast their performance on two examples of different sizes. Both Dengue importation probability and Aedes aegypti mosquito suitability examples show that the methods generate comparatively the same results and the novel analytic method we introduce is dramatically faster than the bootstrap method. 5 - A Multi-attribute Optimization Approach for Post-acute Provider Selection Ashkan Hassani, Texas A&M.University, 4050 ETB, College Station, TX, 77840, United States, ashkanhassani@tamu.edu, Hossein Badri, Maryam Khatami, Mark Alan Lawley In this study a decision making approach is developed for Post-Acute Care Provider (PACP) selection problem. This approach includes two stages. In the first stage, some quality metrics are used in a TOPSIS method to calculate the closeness coefficients of each candidate PACP to short-stay patient’s and long-stay patients. In the second stage, these coefficients are used in a mathematical programming model, along with some other cost and service level metrics to select the best PACP’s for both short-stay and long-stay patients. The proposed approach is implemented for PACP selection problem in the city of Houston, TX, and the results are presented and analyzed. 6 - A Multi-dimensional Evolving Contact Network Approach for Modeling Dynamics of Contagious Disease Spread Buyannemekh Munkhbat, University of Massachusetts-Amherst, 165E Brittany Manor Drive, Amherst, MA, 01002, United States, bmunkhbat@umass.edu, Chaitra Gopalappa, Hari Balasubramanian, Dylan Shepardson, Song Gao Current simulation models for outbreaks of re-emerging infectious diseases (EIDs) take an individual-level or a population-level approach to modeling contact dynamics. Each modeling technique is a trade-off between increased accuracy versus computational efficiency. We propose a novel Multi-dimensional Evolving Contact Network (MECN) algorithm that combines the attributes of both simulation modeling techniques. Considering the uncertainty in type and origin of future EIDs, our proposed methodology will be suitable for analyses of a range of potential EIDs for an integrated approach to disease prevention in low- and middle- income countries. 360A Internet of Manufacturing Things Invited: Invited OR and Advanced Manufacturing Invited Session Chair: Hui Yang, Pennsylvania State University, University Park, PA, 16801, United States, huy25@psu.edu 1 - The Internet of Additive Manufacturing Things on IBM Watson Gayathri Magie, IBM, Kansas City, MO, United States, gayathri@us.ibm.com, Prahalad Rao Learn about IBM’s Watson Internet of Things platform can connect additive manufacturing to the cloud. A road map for implementation at University of Nebraska-Lincoln to register, manage and control the sensor embedded metal additive manufacturing (AM) machine will be revealed. The intent is to send the layer-by-layer sensor data to the cloud, store the data, and use Watson Analytics to detect and predict build failures (currently 1 in 4 parts typically fail to print). This information is used as a repository for part qualification - leading to the certify-as-you-build paradigm in AM. 2 - Parallel Computing and Network Analytics for Fast Industrial Internet-of-things Machine Information Processing and Condition Monitoring Soundar Kumara, Pennsylvania State University, University Park, PA, 16803-2050, United States, skumara@psu.edu, Chen Kan, Hui Yang Rapid advancement in sensing and communication brings a new wave of Industrial Internet of Things (IIoT) technology. IIoT integrates a large number of sensors for smart and connected monitoring of machine conditions. However, existing approaches are limited in their ability to extract pertinent knowledge about manufacturing operations from large volumes of IIoT data. This paper presents new parallel algorithms for large-scale IIoT machine information processing, network modeling, condition monitoring, and fault diagnosis. Experimental results show that our algorithm efficiently and effectively characterizes variations of machine signatures for network modeling and monitoring. MA42

3 - Outlier Detection Based on Minimum Spanning Tree and its Application to Manufacturing Imtiaz Ahmed, Texas A&M.University, College Station, TX, 77843, United States, imtiazavi@tamu.edu, Yu Ding Outliers are points or clusters of points which lie away from the neighboring points and clusters and inconsistent with the overall pattern of the data. The process of identifying outliers become complicated in absence of the labeled training data. Consequently, a supervised learning cannot be used to learn a rule to classify future observations. Minimum Spanning Tree (MST) is a measure, capable of capturing the relative connectedness of data points/clusters and thus can be utilized to identify the potential outliers of a data set in an unsupervised setting. In this work, an MST based outlier detection algorithm is developed and demonstrated using examples in manufacturing processes. 4 - IIoT Automation or Autonomous Manufacturing Thomas Paral, TE Connectivity Ltd, Baden-Württemberg, Germany, thomas.paral@te.com First of all there is no IIoT or Industry 4.0 protocol yet available. There are many initiatives around the globe all speak about IIoT/ Industry 4.0 but most of them are still talking about automation and OEE improvements through transparency not through intelligent autonomous (digital) guided processes along the value chain including life cycle. 360B Internet of Things: Promises, Challenges and Analytics Sponsored: The INFORMS Section on Practice Sponsored Session Chair: Patricia Neri, SAS Institute, Inc., Cary, NC, 27519, United States, patricia.neri@sas.com Co-Chair: Anna Olecka, Honeywell, St Petersburg, FL, 33703, United States, annao007@gmail.com 1 - Internet of Things: Promises, Challenges and Analytics Patricia Neri, SAS.Institute, Inc., 104 Grandtree Ct., Cary, NC, 27519, United States, patricia.neri@sas.com IoT is a technological innovation taking place in key areas: healthcare, transportation, energy production, smart cities and connected factories to name just a few. IoT is possible because of the decreasing cost of sensors and storage, the Cloud and Analytics. Connected devices gather vast amounts of data, which is analyzed, optimized and turned into insight and information. IoT technology is in early development and multiple challenges must be addressed: data availability and ownership, resistance to adoption, data privacy and data security. In this session we will discuss these challenges and specific IoT applications, as well as, jobs that are being created by IoT and the skills required for those jobs. 2 - Panelist Bill Groves, Honeywell, Morris Plains, NJ, United States, William.Groves@Honeywell.com MA43

MA44

360C Behavioral Operations Contributed Session Chair: Matthew Walsman, Rutgers Business School, Berkeley Heights,

NJ, United States, mwalsman@business.rutgers.edu 1 - Decision Support for Airport Surface Operations: A Sociotechnical Systems Approach

Elizabeth M. Argyle, Research Fellow, University of Nottingham, Nottingham, United Kingdom, elizabeth.argyle@nottingham.ac.uk, Jason Atkin, Geert De Maere, Robert J. Houghton, Terry Moore, Herve P. Morvan Demand for airspace capacity is increasing, placing additional pressures both on physical infrastructure and human resources. We present a sociotechnical systems approach for developing a decision support system (DSS) to assist the ground controllers with the complex scheduling and routing tasks at a major European airport. We discuss the role of human behavioral research during the development of search-based algorithms for taxiway routing and departure scheduling and consider the insights that this research has provided into how human operators approach these tasks.

147