Improving Loops that Matter

Abstract

The Loops that Matter (LTM) approach to understanding behavior has proven easy to use and broadly applicable, but it has a shortcoming in its original formulation. This is because the original formulation treats the impact of a flow on a stock relative to the net flow, so that all scores tend to get very large in magnitude as a stock approaches equilibrium, but how big depends strongly on how the flows are specified. By reformulating the link scores from a flow to a stock, this topological dependency is removed. The mathematics behind this approach makes clear the relationship of LTM to the Pathway Participation and Loop Impact analysis methods. The result of this, when applying the analysis to a variety of models, is that the determination of the structure responsible for behavior is clearer, and more clearly tied to work already documented using other techniques.

Measuring the change in behavior of a system with a single metric

“…This is a demonstrably unique measure which has a well-defined relationship to the Loop Impact metric of Hayward and Boswell (2014). For further explanation of the relationships between LTM, Pathway Participation Metric and PPM, see Schoenberg et al (2023).…”

“…LTM computes three metrics which are used to measure loop dominance. These three metrics were first elaborated in Schoenberg et al (2020) and were subsequently updated in Schoenberg et al (2023) (the update first appearing in Stella Architect version 2.1). The three metrics are the link score, the loop score and the relative loop score.…”

“…z is written in a discontinuous form to reflect the implementation of the calculation. Schoenberg et al (2023) shows how to measure the link score between a flow and a stock using the same process.…”

“…Loops that Matter (LTM) is an algorithmic method for performing a “formal assessment of dominant structure and behavior” (Duggan and Oliva, 2013) applicable to models of any size and complexity (Schoenberg, 2020; Schoenberg et al ., 2020, 2023). LTM is the first of the algorithmic methods, developed to perform loop dominance analysis, that has been integrated within commercially available system dynamics software (Stella Architect v2.0 and later) and is, therefore, currently applied by practitioners as well as researchers in the field, including Aboah and Enahoro (2022), Aboah and Setsoafia (2022); Hovmand et al .…”

“…(2020) and were subsequently updated in Schoenberg et al . (2023) (the update first appearing in Stella Architect version 2.1). The three metrics are the link score, the loop score and the relative loop score.…”