IJCAI is recognizing Sandholm "for his contributions to computational economics and the theory and practice of negotiation and coalition formation among computationally bounded agents." Sandholm works at the convergence of AI, computer science, economics, and operations research, focusing on constructing electronic marketplaces that are efficient in terms of economic results and computational processes. He develops systems that lead to economically desirable outcomes on a global scale, despite the fact that the participating agents--humans or software agents--act in their own self-interest. While this has also been a key goal in game theory, Sandholm's work goes beyond traditional game theory because it takes into account the real-world constraints of limited computing power and communication bandwidth, as well as the desire to preserve privacy.
"Tuomas has developed computer systems and concepts for doing multi-dimensional auctions, which play off of different aspects of a product in more creative ways," said James H. Morris, dean of the School of Computer Science (SCS) at Carnegie Mellon. "This award recognizes that his work meets high standards in an important line of research that the experts believe will be increasingly important to the field of AI." Sandholm directs the Agent-Mediated Electronic Marketplaces Laboratory in the Computer Science Department. A native of Helsinki, Finland, he received the equivalent of a combined bachelor's and master's degree in industrial engineering and management science from the Helsinki University of Technology in Finland in 1991. He earned master's and doctor's degrees in computer science from the University of Massachusetts at Amherst, in 1994 and 1996, respectively. Before joining the Carnegie Mellon faculty in 2001, he taught at Washington University in St. Louis from 1996 to 2000 on the assistant and associate professor levels.
Sandholm also works in the industrial arena. He is founder, chairman of the board and chief technology officer of CombineNet, Inc. an optimization software company focusing on procurement issues for major companies. Founded in St. Louis in 2000, CombineNet is now headquartered in Pittsburgh and employs 53 full-time, high tech professionals. According to Sandholm, the company has cleared over $3 billion in procurement auctions for Fortune 500 companies in the last 18 months, saving customers over $300 million.
The savings, Sandholm said, are largely due to a new way of conducting business called expressive competition, which he pioneered. Until now, he said, expressive competition, which leads to vast economic saving for all parties involved, has been impossible because it gives rise to a difficult combinatorial optimization problem revolving around who wins what business. CombineNet has solved this optimization problem. Earlier this year, in recognition of his many achievements, Sandholm was named a Sloan Research Fellow by the Alfred P. Sloan Foundation, which distributes grants to stimulate fundamental research by young scholars of outstanding promise. He is also a recipient of the Inaugural ACM Autonomous Agents Research Award established in 2001 to recognize researchers whose current work is an important influence on the field. He is also the recipient of the National Science Foundation Career Award.
"Sandholm has conducted an intensive research program in electronic markets and multi-agent systems, and has also done work in other areas of autonomous agents," said the ACM citation. "The breadth and depth of his contributions over a relatively short period of time are impressive." In addition to his academic and industrial credentials, Sandholm is a champion windsurfer who was ranked 12th in the world in 1987. He will receive the Computers and Thought award at the IJCAI 2003 conference, which takes place August 7-12 in Acapulco, Mexico. He will receive a $2,000 prize and deliver a plenary lecture on "Making Markets and Democracy Work: A Story of Incentives and Computing."
Sandholm is the second Carnegie Mellon faculty member to receive IJCAI's Computers and Thought award. Computer Science Professor Tom Mitchell received it in 1983 for his work in machine learning.
For more information on Sandholm, see http://www-2.