Informs 2015

116 / 552

INFORMS Philadelphia – 2015
114
4 - A New Approach for Optimal Electricity Planning and Dispatching
with Hourly Time-scale Air Quality
Paul Kerl, Georgia Tech, 755 Ferst Drive, NW, Atlanta, GA,
30332, United States of America,
Valerie Thomas
Energy production from coal, natural gas, oil and biomass generates air pollutants
and health impacts. Pollutant exposure depends on the relative location to power
plants and atmospheric conditions which vary by hour, day and season. We have
developed a method to evaluate pollutant formation from source emissions which
we integrate with an electricity production model. In a case study of Georgia we
show how to reduce health impacts by shifting production during select hourly
periods.
SC63
63-Room 112B, CC
Doing Good with Good OR II
Cluster: Doing Good with Good OR
Invited Session
Chair: Itai Ashlagi, MIT, 100 Main st, Cambridge, Ma, 02139,
United States of America,
Co-Chair: Lisa Maillart, Swanson School of Engineering, Hall
Pittsburgh, PA,
1 - Finding Patterns with a Rotten Core: Data Mining for
Crime Series Detection
Tong Wang, Graduate Student, MIT, 70 Pacific Street, Apt. 242A,
Cambridge, MA, 02139, United States of America,
We worked with the Cambridge Police Department to build a model that can
automatically detect crime series, which analysts spend hours per day doing it
manually. NYPD is currently working with our code, aiming to incorporate it into
a custom software package they are developing which can assist in their daily job.
This project has received widespread media attention.
2 - Infusion Center Process Improvement and Patient
Wait Time Reduction
Mengnan Shen, Georgia Tech, Atlanta, GA, United States of
America, , Xiaoyang Li, Allen Liu,
James Micali, Jisu Park, Yunjie Sun, Emilie Wurmser,
Sung Keun Baek
Winship experienced long wait times and low patient satisfaction. Combining data
analytics, stakeholder interviews, queuing network principles, and detailed
simulation analysis, we improved flow, communication, and visibility throughout
the process. Winship implemented our suggestions, resulting in a 28% reduction
in patient wait times from check-in to chair, a 8.5% increase in patient
satisfaction, and a 6 patients/day increase in throughput.
3 - Using Operations Research to Improve the Health of Patients with
Type 2 Diabetes
Yuanhui Zhang, NC State, United States of America,
We developed OR models for policy evaluation and robust optimization of clinical
regimens for glycemic control for patients with type 2 diabetes. We used the
models to address controversial questions including: whether protocols based on
new medications are more effective than standard regimens. A publication from
this work received substantial press and may help inform treatment
recommendations in the future.
SC64
64-Room 113A, CC
Joint Session DAS/ENRE: Decision Analysis
Applications in Oil and Gas
Sponsor: Decision Analysis & ENRE
Sponsored Session
Chair: Brad Powley, Senior Consultant, Strategic Decisions Group,
745 Emerson Street, Palo Alto, CA, 94301, United States of America,
1 - Defining Prospects for Decision Analysis
Ahren Lacy, Decision Analysis Advisor, Chevron, 1400 Smith St,
#31-128, Houston, TX, 77007, United States of America,
The prospect is the building block of decision analysis. We express our belief
about the likelihood of a prospect’s occurrence (by assigning a probability), and
we express our preference should we obtain it (often using a monetary value
measure). Clarity of action in a complex decision requires careful selection of
distinctions in order to define useful prospects. The author will discuss several
real-world applications in oil and gas projects where clear distinctions led to
clarity of action.
2 - A Probabilistic Analysis of Drilling Strategies in Unconventionals
Robert Hammond, Decision Analyst, Chevron, 1400 Smith St,
Houston, TX, United States of America,
This talk will focus on a probabilistic decision analysis of drilling strategies in an
unconventional oil and gas play that has sporadic areas with low chances of
drilling success. The analysis helped determine the optimal drilling strategy,
including whether to drill multiple wells from a single surface location, which
reduces development costs and environmental footprint, or a spaced well
approach, which in some cases can be used to avoid drilling issues and additional
development costs.
3 - Meta-modeling in Decision Analysis: A Case Study
Brad Powley, Senior Consultant, Strategic Decisions Group, 745
Emerson Street, Palo Alto, CA, 94301, United States of America,
, Eric Bickel
A sophisticated physical model, despite representing an organization’s best
thinking, may be excluded from a decision analysis because it cannot complete a
requisite number of runs in a reasonable amount of time. When facing such a
situation, we created a statistical model of a hydrocarbon reservoir model based
on a handful of previous runs, and used it to conduct a probabilistic simulation on
project economics. This talk introduces that approach, and discusses its merits and
challenges.
4 - A Cognitive Decision Room for High-stakes Decision Making
Jeffrey Kephart, IBM, T. J. Watson Research Center, Yorktown
Heights, NY, 10598, , Debarun Bhattacharjya
We have built a cognitive room in which decision makers use a combination of
speech and gesture to interact with a multi-agent system of decision and
information agents. We overview the hardware and software infrastructure of the
cognitive room, describe a set of interacting decision agents, and illustrate via
several examples how the room enables human decision makers to make better,
more informed decisions in the context of high-stakes decisions in domains such
as mergers and acquisitions.
SC65
65-Room 113B, CC
Systems Engineering and Decision Analysis
Sponsor: Decision Analysis
Sponsored Session
Chair: Robert Bordley, Expert Systems Engr Professional, Booz-Allen-
Hamilton, 525 Choice Court, Troy, Mi, 48085, United States of America,
1 - Limits to Rationality in Systems Engineering
George Hazelrigg, Deputy Division Director, National Science
Foundation, Civil, Mech. & Mfg Innovation,
4201 Wilson Boulevard, Arlington, Va, 22230,
United States of America,
Rationality is a worthwhile goal in any engineering activity enabling optimization
and averting poor choices. But attempts to create a rational framework for
systems engineering fail at the time the second person is assigned to the project.
We outline the limits to rationality in systems engineering and illustrate
consequences. Systems design approaches can have destructive impacts on system
design. This paper presents simple procedures to avoid such problems.
2 - Improving Systems Engineering Trade-Off Studies
Greg Parnell, Professor, University of Arkansas,
Department of Industrial Engineering, Fayetteville, AR, 72701,
United States of America,
Today’s complex systems involve significant uncertainties, multiple stakeholders
with conflicting objectives, and growing affordability concerns. SE trade-offs arise
throughout the system life cycle. Surprisingly many of the published trade-off
studies do not have a strong mathematical foundation, do not provide an
integrated assessment of value and risk, and many do not even consider
uncertainty. We report on a book project to provide best practices using decision
analysis.
3 - Making Product Development Decisions with Decision Analysis
Dennis Buede, President And Executive Director, Innovative
Decisions, 8230 Old Courthouse Road, Suite 460, Vienna, VA,
22182, United States of America,
Formal decision processes during system design are commonly called trade studies
or analyses of alternatives (AoAs). This paper will give an overview of the process
for systems engineering and product development, describe the many kinds of
trade studies that should be undertaken, relate decision analysis to these trade
studies and discuss complexities of system design about which decision analysts
should be aware. Numerous real world examples will be given along the way.
SC63