Learning other agents' preferences in multiagent negotiation
Hung Bui, Dorota H. Kieronska, Svetha Venkatesh · Deakin Research Online (Deakin University) · 1996
In multiagent systems, an agent does not usu-ally have complete information about the pref-erences and decision making processes of other agents. This might prevent the agents from mak-ing coordinated choices, purely due to their ig-norance of what others want. This paper de-scribes the integration of a learning module into a communication-intensive negotiating agent ar-chitecture. The learning module gives the agents the ability to learn about other agents ’ prefer-ences via past interactions. Over time, the agents can incrementally update their models of other agents ’ preferences and use them to make better coordinated decisions. Combining both commu-nication and learning, as two complement knowl-edge acquisition methods, helps to reduce the amount of communication needed on average, and is justified in situations where communica-tion is computationally costly or simply not de-sirable (e.g. to preserve the individual privacy).