Heuristic bidding algorithm using fuzzy neural networks in multiple multi-attribute auctions

This article develops a new heuristic bidding algorithm that guides an agent´s bidding behavior in multiple overlapping auctions for multiple items characterized by multiple attributes. The algorithm operates in the following way. It calculates what it believes are the best set of auctions to bid in. It does this by predicting the auctions closing prices using a fuzzy neural network(FNN), allocating the goods to the customers it is acting on behalf of, and then calculating the satisfaction degree of the allocation. Moreover, as the goods are composed of multiple attributes, the agent may have to make trade-offs between them in its bidding in order to best satisfy the user´s preferences. The use of a fuzzy neural network also allows the decision making criteria of our agent to be adapted to the situation in which it finds itself. We show through empirical evaluation against a number of methods proposed in the multiple auction literature that our bidding strategy performs effectively and robustly in a wide range of scenarios.