TacTex09: Champion of the First Trading Agent Competition on Ad Auctions

Sponsored search [4] is one of the most important forms of Internet advertising available to businesses today. In sponsored search, an advertiser pays to have its advertisement displayed alongside search engine results whenever a user searches for a specific keyword or set of keywords. An advertiser can thereby target only those users who might be interested in the advertiser’s products. Each of the major search engines (Google, Yahoo, and Microsoft) implements sponsored search in a slightly different way, but the overall idea is the same. For each keyword, a keyword auction [5] is run in which advertisers bid an amount that they are willing to pay each time their ad is clicked, and the order in which the ads are displayed is determined by the ranking of the bids (and possibly other factors). The Trading Agent Competition Ad Auctions Game (TAC /AA) [3] was designed to providing a competitive sponsored search environment in which independently created agents can be tested against each other over the course of many simulations. In this report, we describe TacTex, the winning agent in the first (2009) TAC/AA competition. TacTex operates by estimating the full game state from limited information, using these estimates to make predictions, and then optimizing its actions with respect to these predictions. This report is intended to supplement our conference paper on TacTex [6] by providing a more thorough description of the agent. Sections 4 and 5 contain significant added material, while Sections 6, 7, and 8 contain some added details. Algorithms 1 through 11 have been added to summarize certain parts of the agent described in the text. For discussion of the competition and experimental results, see the conference paper.