Designing bidding strategies for trading agents in electronic auctions

TLDR: This paper presents both a formal and a more pragmatical approach for the design of bidding strategies that provide buyer agents with useful heuristic guidelines to participate in auction-based tournaments, and proposes a two-fold method for decision making that does not require the evaluation of the whole set of alternative actions.

Abstract: Auction-based electronic commerce is an increasingly interesting domain for developing trading agents. In this paper we present our first contributions towards the construction of such agents by introducing both a formal and a more pragmatical approach for the design of bidding strategies that provide buyer agents with useful heuristic guidelines to participate in auction-based tournaments. On the one hand, our formal view relies on possibilistic-based decision theory as the means of handling possibilistic uncertainty on the consequences of actions due to the lack of knowledge about the other agents' behaviour. On the other hand for practical reasons we also propose a two-fold method for decision making that does not require the evaluation of the whole set of alternative actions. This approach utilizes global (market-centered) probabilistic information in a first decision step which is subsequently refined by a second decision step based on the individual (rival-centered) possibilistic information induced from the memory of cases composing the history of tournaments. In this way, the resulting bidding strategy balances the agent's short-term benefits, related to the probabilistic information, with its long-term benefits, related to the possibilistic information.