Dec. 17, 1997 Contact Ross West (541) 346-2060 Source: David Etherington (541) 346-0472 EUGENE--Artificial intelligence has beaten the world's top chess player, but can it disentangle the planning and scheduling inefficiencies that cost American business and government billions of dollars per year? The answer is an emphatic `yes,'" according to David Etherington, director of the Computational Intelligence Research Laboratory (CIRL) at the University of Oregon. To help develop this new application of computer power, CIRL has received a $1.1 million grant from the joint Defense Advanced Research Projects Agency and Air Force Rome Laboratories Planning and Decision Aids program. Along with a 1995 grant associated with the same program, CIRL funding for this project has now reached approximately $2.7 million. "Planning and scheduling are critical issues for all kinds of organizations in the public and private sectors. Their needs run the gamut from getting a product built on time at the lowest cost, through disaster recovery, to scheduling airline flights, troop movements, or package delivery," Etherington explains. Artificial intelligence, or AI, has made tremendous progress in the area of planning and scheduling technologies, Etherington says. Researchers are now tackling real-world problems and showing great promise for real economic impact. "There are some kinds of complex problems that machines can address much better than humans can; what we are doing at CIRL is figuring out how to apply these computational abilities to the special challenges of planning and scheduling." The emerging field of "supply chain management," currently growing on the order of 100 percent per year, is developing in response to the need for more efficient planning and scheduling. Etherington notes that companies such as PeopleSoft, i2 Technologies, and SAP are beginning to offer advanced scheduling and planning capabilities. The rapid growth of these companies testifies to the value of these services to the business world, he says. One company expects to provide $50 billion of added value to its customers in the next six years. The new research effort at CIRL will go far beyond what is currently available, Etherington explains. "Most previous research has treated planning as a monolithic process. We are developing new ways of approaching the problem, focusing on breaking the process down into operations conducted by teams with individual responsibilities acting within a collaborative hierarchy. We have proved that it is possible to vastly simplify enormous problems using this approach, but coordinating the hundreds of variables involved in these functions is still a superhuman task, but one well suited to AI." Established in 1993 to conduct basic research in Artificial Intelligence, CIRL has attracted faculty from Stanford University, AT&T Laboratories and the University of Edinburgh. CIRL researchers have developed algorithms that provide the best solutions known to a variety of planning and scheduling problems and have acquired a strong technology transfer orientation. The lab, which is located in the Riverfront Research Park, is supported by external funding sources. -30- #G-7318/Local, OrDailies, OrSci, PDX, OrBUS