System Dynamics vs. agent-based modeling—comparing models and approaches: A literature review and a transformation procedure
(English)Manuscript (preprint) (Other academic)
Systems modeling and simulation methods such as System Dynamics (SD) and agent-based (AB) modeling have been used to foster a better understanding of the dynamics and complexity of natural, technical, and social systems. System Dynamics provides an aggregate-level perspective, highlighting thinking in feedback loops and employing differential equations to model the causal relations in a system, exploring the system's dynamics by numerically solving the equations. Agent-based modeling, in a bottom-up method, focuses on constituent units (agents) and their interactions to explore the emerging behavior at a system level by means of simulation. Comparing these modeling methods can help us understand their strengths and weaknesses in order to choose the right approach for a given modeling problem. It may also support the analysis of a given system to build multiple models using the different approaches and comparing them, in particular to treat fundamental uncertainties in systems modeling and simulation. In this paper, we review the existing studies comparing the SD and AB approaches and models, investigating the aims, methodology, and results of such comparative studies. We also highlight lessons learned for future model comparisons by examining how the corresponding SD and AB models are built for the purpose of comparison. A procedure for transforming System Dynamics models into agent-based models is presented and discussed using examples from the literature.
System Dynamics, agent-based modeling, modeling and simulation, complex systems modeling, individual-based modeling, ordinary differential equations, model comparison, literature review
Other Environmental Engineering
IdentifiersURN: urn:nbn:se:kth:diva-171441OAI: oai:DiVA.org:kth-171441DiVA: diva2:844069
QS 20152015-08-032015-08-032015-08-13Bibliographically approved