University of Minnesota
Software Engineering Center
/

You are here

Representation of Confidence in Assurance Cases using the Beta Distribution

Date of Publication: 
January 2016
Associated Research Groups: 
Abstract: 
Assurance cases are used to document an argument that a system---such as a critical software system---satisfies some desirable property (e.g., safety, security, or reliability). Demonstrating high confidence that the claims made based on an assurance case can be trusted is crucial to the success of the case. Researchers have proposed quantification of confidence as a Baconian probability ratio of eliminated concerns about the assurance case to the total number of identified concerns. In this paper, we extend their work by mapping this discrete ratio to a continuous probability distribution---a beta distribution---enabling different visualizations of the confidence in a claim. Further, the beta distribution allows us to quantify and visualize the uncertainty associated with the expressed confidence. Additionally, by transforming the assurance case into a \textbf{reasoning structure}, we show how confidence calculations can be performed using beta distributions.
Venue: 
Proceedings of the High Assurance Systems Engineering Symposium (accepted for publication) in Orlando, FL, January 2016
bibtex: 
@inproceedings{duan2016, author = "\textbf{Lian Duan} and Sanjai Rayadurgam and Mats Heimdahl and Oleg Sokolsky and Insup Lee", title = "Representation of Confidence in Assurance Cases using the Beta Distribution", booktitle = "Proceedings of the High Assurance Systems Engineering Symposium (accepted for publication)", year = 2016, }