Voting is a process similar in some ways to manufacturing and other domains for applying operations research (OR). The
would be voter has a task and the system aids in task completion with measureable outcomes.
are planning the fall to give our discrete event simulation validated software to the world. The "IZGBS" software
is designed to:
- Predict election lines,
- Generate how many machines are needed so that all can expect
to wait less than X (e.g., 30) minutes with probability greater than 95% at all locations (for officials, other county leaders,
- IZGBS can also allocate a fixed set of resources across multiple locations (for officials).
are collaborating with the Franklin County Board of elections and feel that our software (which can handle check-in and booths
or DREs or almost any type of tandem queue) can help to end in-person waiting times in elections. Yet, we need collaborators
and constructive suggestions. IZGBS stands for Indifference Zone Generalized Binary Search because it can do everything a
generalized binary search can do while allowing declarations to be wrong by a "indifference" parameter.
White Paper on Election Systems
It is easy to underestimate the complexity of voting systems. For example, here in the Columbus area, we have over 370
locations for voting and each location has potentially more than one ballot type. In some elections, certain ballots are twice
as long as others. This could even happen for precincts meeting together.
We Focus on Poll Access/Efficiency/Waiting Lines
The election system is run by legislators and local officials who
have several tasks and challenges.
Challenges include: (1) registration issues, (2) worker recruitment and training,
(3) legal issues, and (4) poll access and/or waiting lines.
The technologies that I have helped create focus primarily on efficiency and
poll access/waiting line reductions. We model the system much like a manufacturing operation with jobs to do (voting) and
resources including voters, machines, poll workers, and/or booths.
Main Technologies and Publications
This data is referenced in the following works. Please cite as appropriate.
- Allen, T. T. (2011), Introduction
to Discrete Event Simulation Theory and Agent-Based Modeling: Voting Systems, Health Care, Military, and Manufacturing, London:
Springer-Verlag. (Textbook teaches discrete event simulation, using election system examples to illustrate all the methods.)
M., M. J. Fry, W. D. Kelton, and T. T. Allen (2014), “Improving Voting Systems through Service-Operations Management,”
Production and Operations Management. (Explains multiple approaches for election officials to reduce lines with minimal additional
expenses, including using our modeling and optimization approaches for deciding how many machines are needed and allocating
them to locations accounting for variable ballot lengths.)
- Yang, M., T. T. Allen, M. J. Fry, and W. D. Kelton (2013),
“The Call for Equity: Simulation-Optimization Models to Minimize the Range of Waiting Times,” IIE Transactions
45(7): 781-795. (Focuses on the simulation optimization methods for allocating voting machines to locations to minimize the
expected range in waiting times from longest to shortest.)
- Li, J., T. T. Allen, and K. Akah (2013), “Could Simulation
Optimization Have Prevented 2012 Central Florida Election Lines?” in Proceedings of the 2013 Winter Simulation Conference,
R. Pasupathy, S.-H. Kim, A. Tolk, R. Hill, and M. E. Kuhl, eds. (Attempts to recreate the 2012 central Florida general election
via simulation and examine what might have happened if voting booths had been apportioned taking ballot length variation into
- Afful-Dadzie, A., T. T. Allen, A. Raqab, and J. Li (2013), “Sufficiency Model-Action Clarification
for Simulation Optimization Applied to an Election System,” in Proceedings of the 2013 Winter Simulation Conference,
R. Pasupathy, S.-H. Kim, A. Tolk, R. Hill, and M. E. Kuhl, eds. (Explores an innovative modeling and visualization approach
to optimize taking data limitations into account with an election system hypothetical example.)
- Allen, T.T. and M.
Bernshteyn (2006), “Mitigating Voter Wait Times,” Chance: A Magazine of the American Statistical Association 19(4):
25-34. (Shows how regression methods can produce estimates of line lengths and how queuing theory can provide insights into
and validated predictions of poll closing times.)
click here to download central Florida 2012 data
click here to download Franklin County 2004 data