Article ID: | iaor201523981 |
Volume: | 30 |
Issue: | 1-2 |
Start Page Number: | 29 |
End Page Number: | 57 |
Publication Date: | Jan 2014 |
Journal: | System Dynamics Review |
Authors: | Hayward John, Boswell Graeme P |
Keywords: | system dynamics, feedback, epidemiological models |
Quantifying the strengths of feedback loops can bring insight into the way the structure of a system dynamics model helps determine behaviour. This paper proposes the concept of loop impact, using the functional relationship between the second and first derivative from the Pathway Participation Metric method, and describes a numerical method to derive the impacts of feedback loops within a system dynamics model. An algorithm is presented that will identify which loop, or loop combination, explains stock behaviour. The method is applied to the yeast, epidemic and market growth models and compared with previous work. It is shown that the method can deal with loops that change polarity, hidden loops and those that self cancel. A procedure is set out to calculate impacts when loops have junctions and graphical converters. It is anticipated that the method will be adopted by system dynamicists and applied to a broader range of models.