Subject: System Dynamics and Agent-Based Modeling

Posted by Linda Booth Sweeney on 7/25/2003
In Reply To:System Dynamics and Agent-Based Modeling Posted by Dr. Niall Palfreyman on 7/25/2003

Message:

This is great! I have been hoping to see interest in the overlap between system dynamics and other systems-related frameworks such as agent-based modeling.

Niall, I appreciated your thorough description of the appropriateness of Stella vs. Mathematica. I'd like to hear what others think about these differences in practical application.

The difference seems to be in the level of perspective you want to take*. Tim Joy in a separate but related message wrote: "SD is about understanding causes and consequences of change." Another aspect of system dynamics is that its practitioners tend to focus on the aggregate behaviors that result from specific policies or strategy decisions. This is in contrast to agent-based modeling approaches that focus on understanding the collective actions of individual agents.

While Stella (and other SD software packages) might model fish population change over time, agent-based programs such as AgentSheets and StarLogo (see
http://www.media.mit.edu/~mres) model the individual rules and actions of individual agents (typically in decentralized systems) such as termite migration, traffic jam and neighborhood formation. While SD models attempt to reveal the underlying causal structure guiding (or driving) the behavior of systems, agent-based modeling seeks to discover the patterns of behavior that arise from individual operating principles or rules.

Ok, I'm very curious to hear what others think.

Linda


*For a terrific article on this topic, see Wilensky and Resnick's (1999) Thinking in Levels: A Dynamic Systems Perspective to Making Sense of the World (Journal of Science Education and Technology).