IV&V Facility Research Program Results and SARP Results  

Advanced Search
Click here to complete a short survey. The results of this survey will be used to help us improve the research program and this website.

Click here to view research projects that had new research results added in the last 90 days.


Validation of System Safety Using Bayesian Belief Networks
Point of Contact Joanne Dugan
jbd@virginia.edu
Dates October 2001 - March 2005
Problem Excellent analysis techniques exist for safety assessment, including formal (axiomatic methods), testing, fault injection, inspections, etc. However, no one method is appropriate for the entire system, and all are difficult to apply to novel systems. A methodology is needed for combining and evaluating qualitative and quantitative evidence of safety gathered via different techniques.
Objective Our objective is to develop a methodology to validate the safety of a safety-critical computer-based system through the "safety case" approach. A "safety case" is built by collecting and evaluating information related to V&V activities. We propose using Bayesian belief networks (BBN) to create an evaluation process that is both qualitative and quantitative.
Results SAS 05 Executive Presentation.ppt
SAS 05 Technical Presentation.ppt
Briefing-Presentation of Results.pdf
Keywords system safety, bayesian belief network, fault injection, safety-critical, system designs
Categories Design Analysis
Domain-Specific Analysis
Formal Methods
Requirements Analysis