 |
     |
|
|
||
|
|
|
|||||||||||||||||
|
MetaNEOS Initial Press Release Chicago, September 20, 1997. A group of researchers from the metacomputing and mathematical optimization communities has just received funding from the National Science Foundation (NSF) for a project entitled "Metacomputing Environments for Optimization." The aim of the project is to design and implement the computational tools needed to solve very large optimization problems on metacomputers, which are collections of computers---PCs, workstations, and parallel processors---that communicate with each other over networks such as the Internet. Optimization problems arise in many financial, scientific, and engineering applications, wherever it is necessary to optimize some property of the system at hand, subject to restrictions on the allowable states. Examples include airline crew scheduling, determination of molecular conformation, weather forecasting, and optimal traffic flow. The new project has been dubbed "metaNEOS" by its investigators, since it combines elements of the metacomputing research of two members of the team with NEOS, the Network-Enabled Optimization System developed by another part of the team. If successful, the project will increase dramatically the utility of next-generation computers and networks, making it possible to harness the power of networked computers for large-scale problem solving in science and engineering. The project was one of seven interdisciplinary projects funded under a new program at NSF entitled "Challenges in Computational Science." The goal of the program is to promote interaction and synthesis between different areas of computational science. Accordingly, the metaNEOS project has identified several areas of optimization in which typical problems require a huge amount of computing time to solve, and which could benefit from the raw power and comparitively low cost of a metacomputing platform. metaNEOS team members face the challenge of developing effective optimization algorithms that are suited to the metacomputer environment. They must also develop tools that distribute the workload for the metacomputer among its constituent computers, coordinating the different tasks that make up a single problems and ensuring robust and efficient operation of the system. Their work will build on techniques and tools provided by the Globus metacomputing toolkit and Condor cluster computing system. The NSF will provide funds of $1.8M over a three-year period to the group of nine researchers, who are based at the University of Chicago, Argonne National Laboratory, Northwestern University, Columbia University, and the University of Wisconsin. For more information, send email to wright@mcs.anl.gov or see the project home page at http://www.mcs.anl.gov/metaneos/ |
