2015 Informs Annual Meeting

POSTER SESSION

INFORMS Philadelphia – 2015

10 - Advanced Decision-making Procedures in Massive Failure Data Classification Keivan Sadeghzadeh, Northeastern University, 27 Payne Rd, Newton, MA, 02461, United States of America, k.sadeghzadeh@neu.edu In many professional areas, management decision-making process is based on the type and size of data where data classification is a necessary procedure. Massive amount of data in high-dimensions are increasingly accessible from various sources and it has become more difficult to process the streaming data in traditional application approaches. This poster presents advanced procedures to analyze high-dimensional failure data in order to facilitate decision-making through data classification. 11 - Exploring Residents Attitude Towards Solar Photovoltaic System Adoption in China Yaqin Sun, Drexel University, 38 Clarence Avenue, Bridgewater, PA, United States of America, ys523@drexel.edu, Xiangrong Liu The research aimed to identify the drivers and dynamics that most encourage Chinese customers to install solar PV systems (SPS) in their residential buildings. A survey was designed and conducted among Chinese residents. The first hand data indicated the importance of increasing awareness of SPS among potential consumers. This research also assessed the impacts of gender on their knowledge of, concerns, and attitudes towards PV adoption. However, no significant difference among gender was found. 12 - Design of Financial Incentive Programs to Promote Net Zero Energy Buildings Promoting net zero energy buildings (NZEB) is among key carbon emissions reduction approaches in the U.S. and in the EU countries. We present a mixed integer programming (MIP) model to aid determining the minimum thresholds of financial incentives that would spur growth in NZEBs. Several combinations of production tax credit and loan interest rates have been investigated for different commercial buildings in Tampa, FL. The results indicate the threshold values of the incentive program parameters. 13 - Multi-objective Scenario Discovery for Climate Change Adaptation Julie Shortridge, PhD Student, Johns Hopkins University, 3400 N. Charles St., Ames Hall 317, Baltimore, MD, 21218, United States of America, julieshortridge@gmail.com, Seth Guikema New methods for decision support under non-probabilistic uncertainty are becoming increasingly popular in the climate change adaptation field. Scenario discovery, as part of the robust decision making framework, uses machine learning to identify multivariate scenarios where a plan or system will perform poorly. In this work, we evaluate different methods for incorporating multiple criteria into the scenario discovery process to assess whether the method used impacts the scenarios identified. 14 - The Unit Commitment Model for Power Interruption Contracts Lakshmi Palaparambil Dinesh, PhD Candidate, University of Cincinnati, 221 Piedmont Avenue Apt. 21, Cincinnati, OH, 45219, United States of America, lakshmi603@gmail.com, Jeffrey Camm The term unit commitment implies which power generation units should be turned on or off in a power plant . When the demand for power is high, power could either be bought from the spot market or the customers could be interrupted using a contract. The problem deals with choosing the right set of customers for interruption using a technique called conjoint optimization and hence reducing the overall costs for the supplier. 15 - Virtual Metrology for Copper Clad Laminate Manufacturing Misuk Kim, Seoul National University, 39-339, Gwanak-ro, Gwanak-gu, Seoul, Korea, Republic of, misuke88@naver.com Virtual metrology predicts wafer quality properties based on sensor values of the equipment in semiconductor manufacturing. It reduces the cost associated with physical metrology as well as identifies important equipment sensor values. We applied it to copper clad laminate for printed circuit board with data from a Korean manufacturer. We not only obtained prediction models with a high accuracy, but also found a number of important, yet previously unknown to engineers, equipment sensors. 16 - Goodness of Fittest for Multinomial Model with Clustered Data Zhiheng Xie, PhD Candidate, University of Kentucky, Lexington, KY, 40503, United States of America, zhiheng.xie@uky.edu Discrete-time Markov chains have been used to analyze the transition of subjects from intact cognition to dementia with transient states, and death as competing risk. We proposed a modified chi-square test statistic which can deal with the clustering effects for the multinomial assumption. We showed our new statistic has a better type I error control when clustering effects presents. We apply the test to the data from the Nun Study, a cohort of 461 participants. Alireza Ghalebani, University of South Florida, Tampa, FL, United States of America, alireza@mail.usf.edu, Tapas Das

17 - Discrete Event Dynamic Simulation for Modeling a Real Job Shop System Golshan Madraki, Ohio University, 15 Station St, Apt. F, Athens, OH, 45701, United States of America, gm705913@ohio.edu A new approach for simulating a job shop system is introduced.The interarrival time of jobs,processing time of machines,time between failures,repair time have general distribution. Previous models consider these parameters deterministic or exponentially distributed. we facilitate estimation of maximum production rate where Buffers capacity,Number of machines in each shop,Number of Lift-truck are efficient 18 - Optimization of Food Production (Ready-To-Eat Meat Sticks) Rebecca Brusky, Data Science Student, University of Nebraska Omaha, 3602 Lincoln Blvd, Omaha, NE, 68131, United States of America, rbrusky@unomaha.edu, Betty Love In the production of ready-to-eat meat sticks, the bottlenecks (dependencies) need to be minimized and number of sticks produced needs to be maximized. Dependent components include equipment flow constraints, smoke room duration and cleaning downtime. The largest downtime factor is the required four-hour cleaning when switching to a non-compatible flavor. This poster documents how a six-flavor production line governed by a set of flavor ordering rules and production demands can be optimized. 19 - Rethinking Principal Component Analysis in EEG Classification Xiaoxia Li, North Dakota State University, 124 East Bison Court, Fargo, ND, 58108-6050, United States of America, xiaoxia.li@ndsu.edu Principal Component Analysis (PCA) is considered to be a powerful tool in dimension reduction. However, it is worth thinking of the suitability of application for EEG signal data. Two EEG datasets collected from alcoholic and control groups were used to test the prediction accuracy before and after PCA transformation with SVM and KNN methods. Based on the classification results, we found that PCA is not valid in EEG signal processing. We also concern that other factors might be confounding. 20 - Strategic Exclusive Supply Contract for Carbon Fiber Reinforced Plastic in the Aviation Industry Kenju Akai, The University of Tokyo, 7-3-1 Hongo, Bunkyo-Ku, Tokyo, Japan, akai@css.t.u-tokyo.ac.jp, Kazuma Sakamoto, Nariaki Nishino, Kazuro Kageyama We investigate the rationality of an exclusive supply contract for Carbon Fiber Reinforced Plastic (CFRP) between Boeing and a Japanese CFRP supplier, Toray. We build a mathematical model of the market for CFRP comprising Toray and the oligopolistic market for aircraft, assuming Airbus, as Boeing’s rival. The subgame perfect Nash equilibria show that both Boeing and Toray obtain the higher profits rather than that in the Cournot Competition. 21 - Hand Motion Identification from Electroencephalography Recordings using Recurrent Neural Network Jinwon An, SNU, 1 Gwanak-ro, Gwanak-gu, Seoul 151-742, Seoul, Korea, Republic of, jinwon@dm.snu.ac.kr, Sungzoon Cho Neurological disabled patients can be aided by brain-computer interface (BCI) prosthetic devices. Grasp and lift tasks are basic actions that needs to be implemented in those devices. In this study, grasp and lift tasks were analyzed by using electroencephalography (EEG) recordings. Various recurrent neural network models were used. It shows that EEG can identify hand motions such as reaching, grasping, loading and retracting with high accuracy. 22 - On Optimization of Carbon Capture, Utilization, and Storage Supply Chains under Uncertainty Mahnaz Asghari, Virginia Tech, 1406 University City Blvd., Blacksburg, VA, 24060, United States of America, mahnaz@vt.edu, Hamed Shakouri Ganjavi Carbon capture, utilization, and storage (CCUS) is a crucial technology to mitigate climate change. Due to the high costs of the technology, a great deal of attention has been focused on how the captured CO2 can be optimally utilized or stored. We study optimizing CCUS supply chains under uncertain environment. In this poster, we present an algorithm to generate a candidate network for CO2 transportation and a model for optimizing the utilization and storage of the captured CO2 in CCUS systems. 23 - On Two-row Chvatal Gomory Cuts

Babak Badri Koohi, Doctoral Student, Virginia Tech, 1406 University City Blvd., Blacksburg, VA, 24060, United States of America, babakbk@vt.edu, Diego Moran

Chvatal-Gomory (CG) cuts are a very important class of cutting planes for solving mixed-integer programs. CG cuts for a polyhedron P are obtained by computing integer hulls of its 1-row relaxations. We study 2-row CG cuts, a generalization of CG cuts that are obtained by computing integer hulls of 2-row relaxations of P. In this poster, we present some basic properties of 2-row CG cuts and discuss their relation to other well-known classes of cuts such as split cuts and (crooked) cross cuts.

200