Electrical Engineering and Computer Science

U-M Wins at 2009 Trading Agent Competition; Introduces New Ad Auctions Game   Bookmark and Share

The Trading Agent Competition (TAC) is an annual international forum designed to promote and encourage high quality research into the trading agent problem. Prof. Michael Wellman led the U-M team that organized the first TAC tournament, based on a travel-shopping scenario, in 2000. Today, TAC includes several challenging market games at each annual event.

At this year's competition in Pasadena, CA, the U-M team ("DeepMaize") led by Prof. Wellman competed in the Supply Chain Management game, winning the competition for the second year in a row. This game simulates a dynamic supply chain environment where agents compete to secure customer orders and components required for production of these orders. The game captures many of the complexities of actual supply chains, where both demand and supply fluctuate and each manufacturer has limited production capacity.

TAC 2009 also featured a new Ad Auctions game, designed by Prof. Wellman and Ph.D. student Patrick Jordan, and developed at U-M by a team of graduate and undergraduate researchers. In this scenario, agents representing Internet advertisers bid for search-engine ad placement over a range of interrelated keyword combinations. A back-end search-user model translates placement over each simulated day to impressions, clicks, and sale conversions, yielding revenue for the advertiser. This premier run of TAC Ad Auctions attracted 15 entrants from eight countries, with the team from the University of Texas taking first place.

Read More:

Trading Agent Competition (TAC) Website
TAC Ad Auction Website