NEOS: Optimization on the Internet

The Internet has been much in the news for the past few years, particularly since the World Wide Web exploded into the popular consciousness in 1994-95. Companies now routinely include their URLs in TV advertisements, billboards, and on the sides of trucks. The NBA displays its \nba.com" in huge letters at courtside at the United Center, home of the Chicago Bulls. For many of us, the Web has become the rst place to turn to for every kind of useful information|and lots of not-so-useful information as well. We routinely use the Web to check weather forecasts and movie listings, obtain driving directions, and get up-to-the-minute golf or basketball scores. One of us recently cancelled our New York Times subscription, choosing instead to read the free, enhanced Web version and to save the trouble of stuung recycling bins. Certainly the Web has changed our personal lives, but what of our professional lives? What kinds of information on the Web would be useful for OR professionals? For academics in the eld, what kinds of educational tools are appropriate? Does it make sense to provide computational services for optimization and OR through the Internet to complement these information/education re-sources? What of Intranets, the intra-organizational networks of computers that are proving to be a major source of revenue for providers of Web technology? To explore these issues, our group at Argonne National Laboratory and Northwestern University launched the NEOS project in late 1994 (NEOS=\Network-Enabled Optimization System"). For the most part, our background was at the theoretical end of the OR spectrum| the analysis, development, and software implementation of optimization algorithms|and we wanted to make stronger and more eeective connections with users of optimization technology, in engineering and basic science as well as in more familiar OR applications. We wanted to give users all the information they need to formulate their problems correctly and to choose the right pieces of software for solving them. We wanted also to give users ready access to a wide collection of state-of-the-art optimization software, without subjecting them to the delays associated with buying software and hardware. We saw the Internet (and its favorite son, the WorldWide Web) as the ideal vehicle for achieving all these aims. Information about optimization algorithms and software is dynamic by nature|it changes on a fairly short time scale|and the Web is the obvious way to deliver such information since it is easily accessed and …