NEOS provides state-of-the-art optimization software over Internet

July 17, 2007

ARGONNE, Ill. (July 17, 2007) – Optimization technology is essential to engineers, scientists, businesses and even students. Optimization programs provide the answers to questions ranging from creating the most efficient allocation of resources to finding the healthiest meal at a fast food restaurant. But writing program after program to solve one optimization problem after another or even using an existing optimization library is often time-consuming or tedious.

To address this issue, researchers at the U.S. Department of Energy's Argonne National Laboratory – in conjunction with researchers at Northwestern University– developed the Network-Enabled Optimization System (NEOS), which provides state-of-the-art numerical optimization software over the Internet. The system has become such a success that it currently leads the field in the number of users at Department of Energy labs.

“Because of its ease of use, NEOS has rapidly gained popularity,” said Jorge Moré, Argonne senior computational mathematician, who has guided the development of NEOS since its inception almost a decade ago. In 1999, fewer than 20,000 people used NEOS. In the last year alone, that number had increased to more than 225,000 users. NEOS is now the premier source of optimization technology on the Web for users of optimization software.

The popularity is due to an increased awareness that many interesting and difficult problems in science and engineering can be formulated and solved as optimization problems, coupled with the fact that more than 50 solvers are now available through the NEOS Server.

NEOS has been used extensively for a variety of applications, including modeling electricity markets, scheduling shifts for airport baggage handlers, predicting global protein folding, studying the brain's representation system and designing low-power VLSI (very-large-scale integration systems).

Here's how the NEOS Server works. A user can select a program, or “solver,” for the type of optimization problem with which he or she is confronted. In order to help with the choice, an “optimization tree” provides suggestions, and each solver comes with sample problems and background information. Once the user defines the problem and selects the solver, it compiles all subroutines, links with the appropriate libraries and does the necessary computations. The user is then sent the solution, along with various runtime statistics.

The latest release of NEOS provides several new features to make the server even more attractive. One new feature is the introduction of a database that enables reliable tracking of available solvers, running processes and jobs waiting to be scheduled.

“It has made a big difference,” said Jason Sarich, the NEOS administrator. “We are spending less time on process management and load-balancing issues and more time on further improvements.”

Another new feature is the XML-RPC (Extensible Markup Language - Remote Procedure Call) interface.

“This gives users considerable flexibility in accessing the system,” said Todd Munson, an assistant computational scientist at Argonne who has been a lead developer of the vast infrastructure behind the NEOS Server. “Users can now easily write their own clients to access the NEOS server through a variety of programming languages,” he said.

The NEOS team consisting of Moré and three Argonne colleagues won the prestigious Beale-Orchard-Hays Prize in 2003 for excellence in computational mathematical programming. In presenting this prize, which is awarded only once every three years, the Mathematical Programming Society noted, “The NEOS Server has had a tremendous impact in the field of optimization. The influence of NEOS is such that in many applied fields the NEOS Server is synonymous with optimization.”

The NEOS Server is maintained under the aegis of the Optimization Technology Center , a joint project of Argonne National Laboratory and Northwestern University , and is operated with support from the Mathematical, Information and Computational Sciences Division of the Office of Advanced Scientific Computing Research in the U.S. Department of Energy's Office of Science.

With employees from more than 60 nations, Argonne National Laboratory brings the world's brightest scientists and engineers together to find exciting and creative new solutions to pressing national problems in science and technology. The nation's first national laboratory, Argonne conducts leading-edge basic and applied scientific research in virtually every scientific discipline. Argonne researchers work closely with researchers from hundreds of companies, universities, and federal, state and municipal agencies to help them solve their specific problems, advance America 's scientific leadership and prepare the nation for a better future. Argonne is managed by UChicago Argonne, LLC for the U.S. Department of Energy's Office of Science.