Skip to Main Content
Agent-based modeling is one of the popular tools for analyzing complex socio-technical systems. Because of the complex nature of such systems a systematic methodology is required to guide the modeling process. By studying the existing methodologies in MAS we distinguished four major differences between MAS and ABM regarding goals, system scale and diversity, level of system understanding and verification and validation concerns. In this paper we take these differences into account and based on more than 25 case studies, we present a methodological framework for developing agent-based models that consists of five general iterative phases. These phases namely: system analysis, model design, detailed design, implementation and evaluation further consists of smaller step that are also addressed in this paper. This methodology provides a tool independent template while respecting the specific requirements for ABM.
Date of Conference: 11-13 April 2011