Distributed group formation among agents based on decisional structures and linguistic mediation rules
Rafik Hadfi, Takayuki Itō · 2012
Autonomous agents operating in a multi-agent environment may seek collaboration in order to fulfill their tasks and maximize their individual payoffs. One way to deal with this situation is to structure the agents as groups whenever the group performance is more efficient than the performance of a single agent. There are several ways to create such groups, considering the environment, the level of cooperation, etc. In this paper, we propose a way to group agents through a set of metrics used to compare the agents’ constraints, utilities as well as the certainties over their possible outcomes. Precisely, we define a decision structure value which gives a reasonable condition under which agents having similar decision structures can form a group. Therefore, formed groups will benefit from the cooperation of its members by satisfying their constraints as well as maximizing their payoffs. Under such criterion, the mediator will perform a clustering of the structural values to form groups and then use a linguistically expressed mediation rule based on ordered weighted averaging operators (OWA) that formalize the consensus policy. Experiments have been performed to test the existence of the decision structure value as well as its capability to describe an agent’s decision structure. Moreover, the decision structure value was used for a distributed hierarchical clustering of the agents.