The need for speed pushes optimization researcher
Posted April 16, 2007
As Todd Munson writes computer code and develops algorithms – mathematical recipes – he gets the greatest kick out of making his programs run faster.
Im a speed junkie, admits Munson, a computer scientist at the Department of Energys Argonne National Laboratory who focuses on optimization research.
Optimization is really a matter of determining what is best for some system, he says. For example, engineers designing a detection system might need to know where to place a set of sensors for optimal coverage, he says. Or, If Im a business, what strategy do I use to get the best profit out of my constituents?.
Despite myriad applications for his work, Munson is driven by the mathematical nuts and bolts of the problems themselves. As a generalist, much of his work is focused on the Toolkit for Advanced Optimization (TAO), an Argonne-based project to develop computer code researchers can apply to a variety of high-level problems. He and the other computer scientists on the project post their code on the Internet, which allows researchers to download and adapt it to meet their specific goals.
For particularly sticky applications or on larger projects supported by the Department of Energys Scientific Discovery through Advanced Computing (SciDAC) program, Munson works more directly with researchers, including economists, chemists, physicists and environmental scientists.