“How can we accumulate knowledge produced by multi-agent social simulations?” This is a central question among researchers using Multi-Agent Based Simulations (MABS) for analyzing social systems. This issue is directly linked to the issue of evaluation and comparison of MABS models: once it is possible to see how two models can interact with, confirm, or contradict each other, it is possible to evaluate and express clearly what each simulation model adds to a body of knowledge.
At the moment, all too often models are very complex, ad hoc, and stand on their own, which makes comparison and reuse in further analysis difficult. Only a few researchers really care about the transferability of their artificial world, explaining clearly the limits of applicability of their model, and the context in which structural properties can be considered as true. The M2M meetings are intended to exploit the power of comparative methodology to advance computational social science.