A Framework for Autonomic Networked Auctions

Business companies are showing a growing interest towards the use of the Internet as part of their business. In particular, networked auctions exploit characteristics that are particularly attractive, and are therefore gaining interest as mean for business. This is especially true when the management efforts, to handle such auctions, are reduced by the use of automatic trading agents. The automatic nature of agents, however, does not allow them to account for the non-stationariness of the market, with the result of inevitably limitating optimization of auction economic factors and, finally, the overall utility. We outline a platform where trading agents behave in an autonomic fashion. Agents employ biologically-inspired techniques, to adapt auctioning strategies to present environmental conditions, through exploitation of self-* properties. We put particular emphasis on self-configuration and selfadaptation, presenting results of preliminary experimentations, and showing through simulations how autonomic behaviour is expected to impact current results.