Adaptive techniques for online auctions

Online auctions consider a setting in which agents arrive dynamically and require that an allocation and payment decision is made before they depart. An auctioneer must take a decision having no knowledge about the time horizon and the strategy of the bidder agent. In this work, we present an utility theory based model which is capable of taking the auctioneer decision and adopt its assumptions about the bidders from the results of past auctions. Simulation results are presented in order to test the efficiency of the proposed model.