Social interaction in artificial society

Recently, the technologies deriving from artificial intelligence and theories of self-organising adaptive systems have furnished new types of models such us cellular automata, multiagent systems and genetic algorithms in order to define simulation in the field of social science. The starting point in these types of models is the individual. The explanation of phenomena is based on a theory of the behaviour of each agent and his relationships with other agents. Therefore, these models have proven to be very useful in the study of social interaction. Actually, computational models that study social interaction use the cognitivist paradigm of bounded rationality introduced by Simon. The concept of bounded rationality provides us with a new perspective to construct sociological theories based on methodological cognitivism. According to this perspective a computational model aims to reproduce social interaction thought conceptual terms and the transformation of an empirical phenomenon in symbolic representations, governed by formal rules that should be consistent with that empirical phenomenon. In this sense, this type of simulation definition is formulated by selecting an "informational" observation level, where the "immateriality" of an empirical phenomenon is preferred. Thus it is impossible to reproduce all the aspects of a social interaction and this leads to a lack of generality that makes it very difficult to compare the results of a simulation with the empirical world