What is System Dynamics?

A short answer is that system dynamics is a "...methodology for studying and managing complex feedback systems, such as one finds in business and other social systems." (System Dynamics Society)

Periodically, this page will be updated with links that add more details to that answer, provide examples and case studies, or illustrate the diverse range of applications within the field of system dynamics. A more thorough description of system dynamics is available from the System Dynamics Group at the University of Bergen.

I like to emphasize that System Dynamics is a Policy-Design Discipline.

Often, a policy issue stems from problematic behavior that unfolds over time. In that context, system dynamics provides an effective method for thinking about the issue in ways that permit both conceptual and simulation modeling of the problem, and new policy options can be identified and evaluated.

The SD policy-design approach is an iterative process that includes these steps:

• identify and graph problematic behavior patterns over a meaningful time horizon,

• identify the main stocks (accumulations) in the system and their corresponding inflows and outflows,

• sketch and describe a conceptual model as a hypothesis for the problematic behavior,

• write equations to transform the conceptual model into a simulatable computer model that can replicate the problematic behavior,

• conduct sensitivity tests, including feedback loop analysis, to determine how the model's structure is contributing to its behavior,

• experiment with policy parameters and feedback structure to identify potentially effective and robust policies for alleviating the problem,

• repeat the iterative process as necessary to improve problem conceputalization, model formulation, and policy design.

Throughout the process, SD practitioners interact with the real-world decision makers who confront the policy issue, drawing on their experience and insight during the conceptualization stage and their judgment during the policy-design stage.

Why the need for this approach?
Policy analysis and design always proceed on the basis of some model of the way the real world generates policy problems. Typically, such models are structures within human memory--mental models--that represent the cumulative experience and insight of those living with or studying the problem. Policy options are then evaluated by mental simulations of expected cause-and-effect relationships.

Persistent policy issues, however, typically arise in systems containing complex information feedback structures and nonlinear relationships, and it is extremely difficult to mentally simulate expected effects. Moreover, when diverse members of a team of analysts engage their individual mental models, the likelihood that they will share the same image of an unfolding process is slim. Any policy design compromise in that context probably depends more on the relative persuasive skills of the analysts than on the relative quality of their mental models.

Useful for policy analysts, even without computers.
Professional education in policy analysis could be enhanced by the addition of systems thinking and system dynamics to the curriculum. Many policy issues can be modeled formally, and many students of the policy process are attracted to formal modeling.

Even when an issue is seemingly intractable and a useful simulation model is beyond grasp, the process of debating alternative dynamic hypotheses in feedback terms can be illuminating. Often, a shared mental model is an important objective within a group of analysts and managers. In that case, a well-developed discipline for thinking systemically can be valuable even to those who do not engage in formal computer modeling.

David Wheat
Senior Lecturer, University of Bergen
President, Wheat Resources Inc.