Using principles from Expected Utility Theory, we analyze the criteria that a customer agent in agent-based automated contracting would use in making decisions during the bidding cycle. We use the University of Minnesota's MAGNET automated-contracting environment as a framework for this analysis. Two decisions must be made by a customer agent during this process: deciding the composition of the Request for Quotes, and evaluating and awarding bids. We show how principles from Expected Utility Theory can be applied in a mixed-initiative environment, where user preferences control the decision-making process, and user decisions are final. Finally, we show how the market infrastructure can support agent decision-making by gathering and analyzing statistical data on activities in the market.
|Original language||English (US)|
|Number of pages||8|
|State||Published - 2000|
|Event||4th International Conference on Autonomous Agents - Barcelona, Spain|
Duration: Jun 3 2000 → Jun 7 2000
|Other||4th International Conference on Autonomous Agents|
|Period||6/3/00 → 6/7/00|