New Scalability Model Ranks Certus as Most Efficient Assurance Method

Researchers at the University of Victoria have developed a scalability model to quantify the decision complexity and effort required for applying quantitative confidence assessment methods to assurance cases—structured arguments used to demonstrate system safety or reliability. The model accounts for both worst-case and average-case scenarios, scaling with argument size. Prior work identified high effort as a barrier to adoption, so the authors parameterized their model using published case study data and tested it on three existing methods: Bayesian Belief Networks (BBN), Dempster-Shafer Theory (DST), and the Certus method.

Results show that while Certus has the highest worst-case decision complexity, its average-case effort is lower than both BBN and DST. The paper provides a practical tool for researchers refining or creating new confidence assessment methods, enabling them to estimate deployment effort early. This trade-off—higher worst-case complexity but better average performance—suggests Certus may be optimal for typical assurance case sizes, though careful planning is needed for edge cases with very large arguments. The preprint is accepted for SafeCOMP'26 Workshop Proceedings.