They may not look like detectives hot on the trail of criminals, but the students and faculty advisers in TU's Applied Mathematics Laboratory (AML) are applying mathematical modeling techniques to one of law-enforcement's knottiest problems.
The lab, based in Stephens Hall, partners with businesses and government agencies to develop mathematical solutions to real-world problems. Furnished with a couple of desktop computers, a small library and chalkboards, it's an unremarkable setting for a remarkable challenge: to develop a method that can examine a serial offender's crimes, then estimate the likely location of that offender's "home base."
"The challenge is to define an optimal police search area," says Michael O'Leary, AML program director and associate professor of mathematics. "If you know where a serial arsonist, burglar or rapist committed crimes, you can theoretically go back and determine that person's base of operations.
Funded by a grant from the National Institute of Justice (NIJ)—the research and development arm of the U.S. Department of Justice—AML's six undergraduates and one grad student will collaborate with area police agencies that gather statistics and report their findings to an NIJ project manager.
O'Leary says his colleague Andrew Engel, an AML adviser, played a key role in bringing the project to TU. "Dr. Engel met Stan Erickson, NIJ's chief of research and technical development, at a professional meeting, and they ended up discussing potential projects.
"The National Institute of Justice has a number of very interesting research questions, so they were happy to find out about the lab," he says. "It's a win for them, for the university and for our students—this is a good project with very real applications."
Under the direction of professors Engel and Coy L. May, the seven AML students—Brooke Belcher, Brandie Biddy, Paul Corbitt, Gregory Emerson, Laurel Mount, Ruozhen Yao and Melissa Zimmerman—will develop mathematical models they hope will enable them to improve on existing models.
Those existing models aren't perfect, says O'Leary. "Right now there are models that say 'working from this set of assumptions, we can make a prediction about an offender's home base,'" he says.
"Our AML students are trying to incorporate algorithms that really work," he adds. "We're looking to develop software that could eventually be used by police departments throughout the country."
Office of Sponsored Programs & Research