Antagonism and Identity
J. Erceau, Ph. Benhamou · World Futures · 1994
Distributed artificial intelligence (DAI) aims to solve problems using cooperation between real or abstract entities. The transition from artificial intelligence (AI) to distributed artificial intelligence poses the problem of transition from an individualistic to a collective concept. In AI, there is no problem of identity (no external references, no differences). In DAI, however, the identity of the agent is defined in accordance with other agents with which it forms a group. This paper attempts to build a model of agent identity taking into account the duality between individuality and collectivity. First, a brief description of the dynamic logic of contradiction and a formalism for agent representation are presented. Finally, a synthesis is proposed in the form of the definition of agent identity seen as a dynamic balance or fixed point for a set of functions composed of antagonistic action pairs.