Studying Dynamic Change Probabilities and Their Role in Change Propagation

Long-lived systems are likely to experience many independent modifications during their lifecycles. Prior literature provides tools for predicting how a change in a fixed system is likely to propagate, but these tools do not address change propagation across multiple, independent modifications. The phenomenon of a modification consuming excess, thereby increasing the likelihood of change propagation in future modifications, is studied in this work as dynamic change probabilities (DCP). This research builds on change propagation techniques, network theory, and excess to provide high-level guidance about how DCP may alter change propagation within a system over time. A sample of existing and synthetic systems are explored, as we show that the rate of change likelihood increase following a modification depends on the number of components (nodes), the dependencies between components (edges), and initial change propagation probability values (edge weights). Results also show that excess placement in specific components, and the presence of system hubs (high-degree components), can mitigate the impact of excess consumption when multiple system modifications are made over time.