EAST LANSING, Mich. -- Andrew Christlieb is all about making things simple – breaking problems into manageable chunks.
Except his problems start out large: mathematical problems so large that "there does not exist a large enough supercomputer in the world to solve them in my lifetime," said Christlieb, a Michigan State University assistant professor of mathematics. "Even with all the horsepower of all the computers in the world, there's not enough power to solve them."
To turn up the power, the U.S. Air Force Office of Science and Research has begun a program aimed at identifying scientists whose research may offer significant advances in key areas. The Young Investigator Research Program will spend $6.3 million on 21 scientists and engineers.
Christlieb has been awarded $300,000 over three years after competing with 144 other proposals in the program's first year.
The program supports scientists and engineers who have received doctorate or equivalent degrees in the last five years. Grant recipients must show exceptional ability and promise for conducting basic research. The objective of this program is to foster creative research in science and engineering, enhance early career development, and increase opportunities to address the Air Force mission and the related challenges in science and engineering.
Christlieb will study questions of plasma physics, or ionized gas, the fourth state of matter (after solid, liquid and gas) that is perhaps most popularly used in flat screen video monitors. Plasmas are also behind much of astrophysics and many energy questions, including fusion, and are crucial in the fabrication of computer chips and a widerange of manufacturing issues. One of Christlieb's favorite applications is the use of plasma as a high-efficiency propellant in next-generation spacecraft propulsion systems.
Plasmas interest Christlieb because "they embody a rich set of complex dynamics that require both careful mathematical analysis and sophisticated numerical methods to obtain an understanding of their full range of possible behaviors."
Those issues, such as reproducing the fusion that fashions the sun in a laboratory, involve tremendously complex calculations that outpace existing computer capabilities, Christlieb explains. Modeling solutions to real devices, such as energy and manufacturing systems, often requires a new way of looking at those problems.
Math is the solution – offering ways to rephrase questions which are optimized for computing. This can dramatically speed up calculations of difficult problems, such as those found in engineering, physics, chemistry and mathematical biology.
"Although advances in computer technology are doubling every two years, many important problems in physics and engineering are out of reach for the foreseeable future unless we change how we approach these problems," Christlieb said.
Christlieb said creating new algorithms – ways to direct a computer to process information – offers the hope of reducing some of these weighty problems to questions that can be answered in manageable amounts of time.
Christlieb, along with his collaborators, work on what he calls "clever math tricks" to create accurate computer models of dynamic plasmas that are designed to test the validity and accuracy of the new methods.
The stakes are high.
Christlieb will work with his colleagues at Air Force Research Lab to transition some of these methods into production scale codes, which are used in the simulations of complex and expensive equipment. Long before research and development offices start cutting metal, a product is built with these computer models. Improving the speed and accuracy of computer models is vastly more cost- effective.
"If you can do the design in the computer, then virtually all you've done is spent the power of running the computers," Christlieb said. "You haven't spent money to cut it, assemble it or the labor to debug the system. Hopefully you've done it all in the computer ahead of time. I try to build a better mousetrap for understanding plasma dynamics so that complex problems can be made more tractable. Hopefully, my work will benefit those who are working on design of complex systems which make use of plasma."
Contact: Sue Nichols, University Relations: (517) 353-8942, Nichols@msu.edu