The necessity of intelligent agents in social simulation

The social simulation field is here argued to show a history of growing complexity, especially at the agent level. The simulation of the emergence of Macro-social phenomena has required heterogeneous and dynamic agents, at least in the sense of agents moving in a physical and social space. In turn, the simulation of learning and evolutionary algorithms allowed for a two-way account of the Micro-Macro link. In this presentation, a third step is envisaged in this process, and a 3-layer representation is outlined: the Micro, Mental, and Macro layers. This complex representation is shown to be necessary for understanding the functioning of social institutions. The 3-layer model is briefly discussed, and specific cognitive structures, which evolved to cooperate with emerging Macro-social systems and institutions, are analysed. Finally, social intelligence is argued to receive growing attention is several fields and applications of the science of artificial, with which social simulation is interfaced or will soon be.