Models, abstractions and phases in multi-agent based simulation

This paper introduces some considerations about simulation practice and a schema of the models that are implicitly and explicitly involved in a Multi-Agent Based Simulation (MABS). The aim of this work is to set simulation inside scientific framework. In order to do that we give an interpretation of the levels that compound a simulation and that constitute different kinds of abstraction. A clear awareness of the relations that exist between these levels and the corresponding steps, in fact, it is necessary if MABS wants to be adopted as a scientific investigation method. Our opinion is that this analysis suggests some answers to the objections that are often directed towards the use of simulation in scientific practice but also underlines some criticalities in this process.