AuthorsS. Nair, N. Walkinshaw, T. Kelly and J. L. de la Vara
TitleAn Evidential Reasoning Approach for Assessing Confidence in Safety Evidence
AfilliationSoftware Engineering
Project(s)No Simula project
StatusPublished
Publication TypeTechnical reports
Year of Publication2014
Number2014-17
PublisherSimula Research Laboratory
Abstract

Safety cases present the arguments and evidence that can be used to justify the acceptable safety of a system. Many secondary factors such as the tools and technique used to create the evidence, and the experience of the evidence creator, can affect the assessor's confidence in the evidence cited by a safety case. One means of reasoning about the confidence established in the evidence is to present an explicit confidence argument that corroborates the reason for having confidence on the evidence. In this paper, we propose a novel approach to automatically construct these confidence arguments through asking assessors to provide individual judgements concerning the trustworthiness of evidence and the appropriateness of its use in supporting the case. These judgements can be supported by further evidence, simply asserted, or expressed with stated uncertainty. The proposed approach enables these judgements to be presented within the context of an overall argument of confidence, and a quantified aggregate of the overall confidence to be derived. The approach is based on Evidential Reasoning - a decision-theoretical technique for reasoning about uncertainty and evidence. Our approach enables assessors to clearly present complex reasoning concerning evidence whilst making any doubt or uncertainty explicit. The proposed approach is supported by a prototype tool (EviCA) and is evaluated using the Technology Acceptance Model.

Citation KeyEviCATechReport