Contact: Maridith Geuder
From helping predict the spread of diseases to planning emergency responses for natural or terror-related disasters and other complex events, physics theories are coming to the aid of sometimes-limited computer technologies.
Mississippi State physics department head Mark A. Novotny in a recent issue of Science Magazine describes one example. Science is the weekly journal of the American Association for the Advancement of Science, the world's largest scientific society.
In the technical article, Novotny describes a theoretical process he designed by which multiple desktop computers linked together can increase the possibility for sorting information and providing data for quick decisions.
Novotny said a $450,000 National Science Foundation grant has supported his research in collaboration with Gyorgy Korniss of Rensselaer Polytechnic Institute in New York. "Our project suggests a solution to the challenge of how we can use computers to model extremely complex, rapidly changing events," he explained.
Novotny said current modeling systems are severely limited in the number of computer processors that can run discrete-event simulations effectively and accurately. In disease simulations, a discrete event is when an individual contracts the disease. In a disaster scenario, a discrete event is when a fire or ambulance unit is dispatched to a particular location, he said.
Novotny's research team focused on developing mathematical algorithms that could optimize simulation speed and data management at the same time. That's a daunting task if one is dealing with events that occur randomly--such as last summer's national West Nile virus outbreak.
In addition to his academic duties, Novotny also heads the ERC Center for Computational Sciences at MSU. His investigation, which involved a number of the ERC's more than 1,000 linked processors, explored how parallel computing could assist in dealing with such problems as the spread of disease.
Called parallel discrete event simulation, or PDES, the approach applies physics to create and control an evolving and fluctuating "virtual" time that can simulate real events. The process represents a major improvement over systems that now can effectively use only five to 10 computers simultaneously.
"We've developed a different approach that allows us to run computer codes on more than 100 processors simultaneously to solve a problem," he said. "It will work on any number of processors and the efficiency won't decrease."
Novotny and Korniss, who say the approach has countless real-world applications, have applied for a patent through MSU on their discovery. The next stage of their project--applying PDES to the issue of first-response situations--will involve collaborations with scientists at the University of Alabama at Birmingham.
"We'll look at how best to quickly provide the authorities information about possible outcomes of their decisions in their responses to multi-location calamities," Novotny said. "Where do you send ambulances? Where do you send people? We believe this work can make a difference on some of the important issues of our day."
NEWS EDITORS/DIRECTORS: For more information about the research, contact Dr. Novotny at (662) 325-2806 or man40@ra.msstate.edu.