Abstract
The methodology of artificial societies, coming from sophisticated development of so called agent based models, is going to be a breakthrough in social sciences. In this paper I try to demonstrate it on the examples of mathematical economics and related topics. The growing complexity of mathematical models can be overcome by construction of artificial worlds and computer experiments. The first non trivial results are on the basis of this methodology are already in place. We may also formulate a number of intriguing issues, the answer to which one can expect under further application of the new methodology.