AUSTIN, Texas—A team from The University of Texas at Austin beat 19 competitors in a computer simulation challenge that precedes an international simulation competition in Mexico.
Using an on-screen “agent” developed with an artificial-intelligence computer program, computer scientist Peter Stone, graduate students David Pardoe and Ronggang Yu, and undergraduate Ali Niaz, placed first in the seeding round of the supply chain management competition at the Trading Agent Competition (TAC). The seeding round was held from July 14-18. Stone’s team demonstrated that their autonomous agent was best at handling the multiple decisions that a computer manufacturer needs to make to respond to changing market conditions.
Their agent, TacTex, and others developed by teams from Harvard, Cornell, the University of Western Sydney and elsewhere competed in more than 100 one-on-one games during the five-day, online event. The supply chain management agents vied to buy disk drives, motherboards and other computer parts, decided what type of computer to produce and sold the computers based on market demands.
This is the first year for the supply chain challenge. It builds upon a previous scenario in which on-screen travel agents haggled to create the best vacation packages based on flight, hotel and entertainment options. In the travel-agent version, now referred to as TAC Classic, Stone finished in first place in 2000 and was one of two first-place winners in the 2001 TAC. He entered the 2001 agent, unchanged, as a benchmark in this year’s seeding round of the TAC Classic, where it finished second.
Stone’s supply chain team will compete in the Trading Agent Competition Finals in Acapulco, Mexico, Aug. 11-13. If they do well, the honors will be added to an early July win at the RoboCup international soccer competition in Padua, Italy. During that July 2-11 event, the assistant professor of computer sciences and three students won an agent-guided computer simulated soccer competition.
U.S. News and World Report ranks the graduate program of the university’s Department of Computer Sciences among the top 10 nationally, and its program in artificial intelligence is number five nationally.
For more information contact: Barbra Rodriguez, College of Natural Sciences, 512-232-0675.
