ABSTRACT
The hybrid system SOCAIN is described as a formal model of the self modeling of social systems. The system consists of a stochastic cellular automaton (CA) coupled with a genetic algorithm (GA) and an interactive net (IN), also coupled with a GA. The IN is understood as the self-model of the CA. It can be shown that the adaptive success of the whole system SOCAIN is much greater than the successes of both subsystems alone. Reasons for this are given and the logic of the system is discussed both logically and sociologically. The program SOCAIN contains a CA, horizontally coupled with an interactive neural net (IN), and GAs operating on both base systems. Hybrid systems offer important possibilities for the modeling and simulating of social systems. The CA was constructed to simulate social differentiation, namely the differentiation of a homogenous society into a class society.
