Nowadays agents' models are widely applied to describe and analyse complex systems as different as animal swarms, social systems, economics, biological systems and many others. The common key idea behind these models is that they exhibit a two levels description: a global one (macro level) as a whole system and a second local behaviour (micro level), each agent obeying to some "personal" rules. The main task is to achieve a description of the macro level as emerging from the micro rules.