2016 INFORMS Annual Meeting Program

486 / 561

INFORMS Nashville – 2016
486
WE07
102B-MCC
Data Mining in Decision Analytics
Sponsored: Data Mining
Sponsored Session
Chair: Roy Jafari Marandi, Mississippi State University, MD, United
States,
1 - Next-generation Sequencing (NGS) Data Analysis: Developing A
Scalable Framework For The Future
Michael Chuang, State University of New York, New Paltz, NY,
12601, United States,
NGS analysis presents a domain for biomedical and information technology
professionals to explore. Due to the large amount of data involved and various
constraints of technologies, we delineate issues to consider to develop a
framework using parallel computing and NoSQL database service to greatly
reduce the required time under less infrastructure investments while achieving
satisfactory accuracy.
2 - The Magnificent 7- The Killer Data Mining Errors
Samuel Koslowsky, Senior Analytic Consultant, Harte Hanks, 2118
Ave T, Brooklyn, NY, 11229, United States,
Most managers agree that data mining plays a critical role in assuring a successful
marketing campaign. At times, errors can creep in. While analytic and technical
errors can certainly harm a data mining exercise, most of the problems that
emerge in a modeling project have little to do with technical issues. Rather, basic
reasoning, and marketing related issues are at fault. Errors emerge from all phases
of an exercise. From establishing an appropriate objective, to allowing sufficient
time for completion, to misinterpreting the results to deploying results incorrectly.
A good data mining analysis requires qualified personnel, domain knowledge
experts, analysts, and IT professionals.
3 - Estimating Distance Decay Functions For Arts & Culture Markets
Young Woong Park, Technical Professor, Southern Methodist
University, 6212 Bishop Blvd. Fincher 303, Dallas, TX, 75275,
United States, Glenn Voss
Distance decay functions capture the effect of distance on interaction intensity.
Unlike typical efforts that use distance as a sole independent variable, we estimate
a model that uses organizational, market, and demographic characteristics to
explain variance across geographic markets. We build the model using transaction
data for 7M HHs in 6 geographic markets and investigate characteristics predicting
decay function shape and market-level differences. The resulting model can
estimate decay functions in the absence of interaction intensity data.
4 - Self Organizing And Error Driven (SOED) Artificial Neural Network
For Smarter Classifications
Ruholla Jafari Marandi, Research Assistant, Mississippi State
University, Starkville, MS, 39759, United States,
Mojtaba Khanzadeh, Brian Smith, Linkan
Bian
Albeit Artificial Neural networks’ high prediction power, the technique suffers
from drawbacks such as intransparency. In this paper, motivated by learning
styles in human brains, ANN’s shortcomings have been assuaged and, its
prediction power has also been improved. Self-Organizing Map and Feedforward
ANN are hybridized to solidify their benefits and help remove their limitations.
The proposed method, which we have named Self-Organizing Error-Driven
(SOED) Artificial Neural network, showed significant improvements in
comparison with usual ANNs. Through experiencing 5 different datasets, we
showed SOED is a more accurate, more reliable and more transparent technique.
WE08
103A-MCC
Improving Electricity Grid Flexibility Under
Uncertainty
Sponsored: Energy, Natural Res & the Environment, Energy I
Electricity
Sponsored Session
Chair: Feng Qiu, Argonne National Laboratory, 9700 S. Cass Avenue,
Lemont, IL, 60439, United States,
1 - Enhancing Flexibility Of Power Systems With Intelligent Periphery
Yunhe Hou, University of Hong Kong,
Flexibility is a critical prerequisite for accommodating large-scale variable
renewables before a clean, efficient, reliable, resilient, and responsive smart grid
can be established. In this talk, the metrics for assessing flexibility of the systems
with large-scale renewable integration will be discussed first. Second, a method
for enhancing flexibility, entitled risk-limiting dispatch, will be introduced.
Finally, the operating strategies of intelligent periphery with electric springs will
be discussed as a powerful tool to enhance flexibility of systems.
2 - Robust Defense Strategy For Gas-electric Systems Against
Malicious Attacks
Cheng Wang, Tsinghua University, Beijing, China,
, Wei Wei, Jianhui Wang, Feng Liu,
Feng Qiu, Shengwei Mei
This talk proposes a methodology to identify and protect vulnerable components
of connected gas and electric infrastructures from malicious attacks, and to
guarantee a resilient and flexible operation by deploying valid corrective actions
(while accounting for the interdependency of gas pipeline network and power
transmission network). The proposed mathematical formulation reduces to a tri-
level optimization problem. By reformulating the lower level problem as a mixed
integer linear programming , a nested column-and-constraint generation
algorithm is developed to solve the min-max-min model. Case studies
demonstrate the effectiveness and efficiency of the proposed methodology.
3 - A Study Of Ramp Management And Its Compensation Schemes
Dane Andrew Schiro, ISO New England, Holyoke, MA, United
States, Eugene Litvinov, Tongxin Zheng, Feng
Zhao
The integration of renewable generation could make it more difficult for U.S. ISOs
to satisfy real-time power balance constraints. This talk will summarize the
existing power balance issue, present the current solution of ISO New England,
and explore two potentially better solutions: ramp products and multi-period
market clearing. Formulations of these new methods will be presented along with
discussions of their foreseeable issues. It is hoped that this talk will encourage
rigorous investigation into these emerging ideas, thus aiding in future ISO market
improvements.
4 - Provide Ramping Service With Wind To Enhance Power System
Operational Flexibility
Qin Wang, National Renewable Energy Laboratory (NREL),
Golden, CO, United States, Bri-Mathias Hodge
Maintaining the power system balance requires controllable resources to adjust
their power output to match the time-varying net load. This is becoming more
challenging when the proportion of generation from variable and uncertain
renewable resources in the system is high. This presentation will demonstrate the
feasibility and approaches to rely on wind power to provide ramping service in
the electricity markets. Advanced wind ramp forecasting methodologies are
discussed. In addition, methods on how to quantify power system flexibility
enhancement by using wind power to provide ramping service will be presented.
WE09
103B-MCC
Logistics of Biomass Feedstock for Liquid Fuel
Production
Sponsored: Energy, Natural Res & the Environment I Environment &
Sustainability
Sponsored Session
Chair: Daniela Gonzales, Texas A & M University, 3014 Jennifer Drive,
College Station, TX, 77845, United States,
1 - A Two-Stage Chance-constrained Stochastic Programming Model
For a Bio-fuel Supply Chain Network With Uncertain Biomass
Supply
Md Abdul Quddus, PhD Student, Mississippi State University,
Department of Industrial & Systems Engineering, PO Box 9542,
Starkville, MS, 39762, United States, , Sudipta
Chowdhury, Mohammad Marufuzzaman
This study presents a two-stage chance-constrained stochastic programming
model that captures the uncertainties due to feedstock seasonality in a bio-fuel
supply chain network. The chance constraint ensures that, with a high
probability, Municipal Solid Waste (MSW) will be utilized for bio-fuel production.
To solve our proposed optimization model, we use a combined sample average
approximation algorithm which is made faster by using star-inequalities. We use
the state of Mississippi as a test bed to visualize and validate the modeling results.
Our computational experiments reveal some insightful results about the impact of
MSW utilization on a bio-fuel supply chain network performance.
WE07