The use of Bayesian belief nets in safety assessment of software based systems

The paper describes an investigation of methods to perform a reliability an d safety assessment of the software in programmable safety relevant systems . It emphasises in particular how disparate information sources and differe nt quantitative and qualitative methods should be combined in such an asses sment. It starts with a general discussion of rule based, probabilistic and expert judgement methods and their applicability on software reliability. Then a method for combining different evidences in a reliability and safety assessment is pinpointed, viz. the Bayesian Belief Net (BBN) methodology. It is also illustrated how this method may be applied for safety assessment of software.