A Structural Functionalist Conception of Artificial Agent Societies

This paper reports a case of the successful application of sociologic theories in the area of Distributed Artificial Intelligence. We argue in favour of a rather deep conceptual transfer of sociologic models: it is shown how analytical theories from the school of structural functionalism, aimed at explaining observed phenomena in human societies, can be put “upside down”, so as to come up with a constructive co-ordination mechanism for artificial agent societies. Setting out from such an engineering perspective, we develop a formal model of this mechanism, sketch its operationalisation and illustrate its dynamics by an example.