Model validity in agent-based simulation

In this talk I'll try to formalize the concept of model validity in social simulations using agent-based models(ABM) from two stand points, the empirical approach and the interpretive approach. In the spirit of mathematical theory of general systems I start with the most general mathematical entity describing "validity" using the homomorphism concept between systems. With this general framework the model validity can be formalized from these two stand points. Then, I try to claim that the validity discussion in ABM can be done by mixing these two validity concepts. Furthermore, I will try to interpret validity discussions in some published ABM papers within this general framework.