Multi-agent negotiation using linguistically expressed mediation rules

The problem of multi-agent negotiation is considered. We provide an framework for the multi-agent negotiation process in which each of the participating agents provides a preference function over the set of alternatives. This framework involves a mediation step in which the individual agent preference functions are aggregated to obtain a group preference function. The determination of the satisfaction of a stopping rule which decides whether a suitable final group preference function has been obtained or whether the agents must participate in another round of mediation. It also involves a selection procedure for choosing a alternative based on the final group preference function. We describe various implementations for these different steps. Considerable interest is focused on the implementation of the mediation rule where we allow for a linguistic description of the rule using fuzzy logic. A particularly notable feature of our approach is the inclusion in the mediation step of a mechanism rewarding the agents for being open to alternatives other then simply their most preferred.