Proceedings of the AAAI Spring Symposium on Adaptation, Co-evolution and Learning in Multiagent Systems, March 25-27, 1996., Stanford University, California, USA
This paper presents an adaptive model for multi-agent coordination based on the metaphor of economic markets. This model has been used to develop SIGMA, a system for filtering Usenet Netnews which is able to cope with the non-stationary and partially observable nature of the information filtering task at hand. SIGMA integrates a number of different learning and adaptation techniques, including reinforcement learning, bidding price adjustment, and relevance feedback. Aspects of these are discussed below.