Argumentation and Multi-Agent Decision Making
Simon Parsons, Nicholas R. Jennings · ePrints Soton (University of Southampton) · 1998
This paper summarises our on-going work on mixedinitiative decision making which extends both classical decision theory and a symbolic theory of decision making based on argumentation to a multi-agent domain. Introduction One focus of our work at Queen Mary and Westfield College is the development of multi-agent systems which deal with real world problems, an example being the diagnosis of faults in electricity distribution networks (Jennings et al. 1996). These systems are mixed-initiative in the sense that they depend upon interactions between agents---no single agent has sufficient skills or resources to carry out the tasks which the multi-agent system as a whole is faced with. Because the systems are built to operate in the real world, the agents are forced to deal with the usual problems of incomplete and uncertain information, and increasingly we are turning to the use of techniques from decision theory, both classical and non-standard, in order to ensure that our agents make so...