FOSLS makes no bones about improving computer models
Posted June 4, 2007
If you’ve ever unknotted a gnarled garden hose to let water flow freely, you may have an elementary feeling for Tom Manteuffel’s job.
Manteuffel, a University of Colorado professor, and his fellow applied mathematicians “unravel” complex equations, making them easier for computers to solve.
The technique, called first-order system least-squares (FOSLS), can help accelerate computer simulations of physical phenomena. In one case, FOSLS, combined with a high-speed solution technique, ran a computer simulation of blood flowing through vessels significantly faster than a commercial computer code.
FOSLS could make faster, better models of cardiovascular systems so medical device companies can do tests more cheaply and quickly than they can when using human or animal subjects. Surgeons could develop better techniques without actually operating.
FOSLS itself, however, is not a computer application, but a tool with broad uses.
“We’re developing a methodology for approaching a wide variety of applications and computing a solution more effectively and efficiently – and that includes not just computer time, but human time,” Manteuffel says.
Mixing up some formulas
It all starts with partial differential equations (PDEs). These complex formulas describe continuous processes that are distributed in space and time, such as fluid flow, electrodynamics, and the propagation of heat or sound.
