Software reliability based on software measures applying Bayesian technique

Abstract

Safety critical systems such as nuclear power plants, chemical plants, avionics, etc., see an increasing usage of computer-based controls in regulation, protection, and control systems. Reliability is an important quality factor for such safety critical digital systems. The characteristics of such digital critical systems are explicitly or implicitly reflected by its software engineering measures. Therefore, these measures can be used to infer or predict the reliability of the system. Hence Software Engineering measures are the best indicators of the software reliability. This paper proposes a methodology to predict software reliability using software measures. The selected measures are used to develop Bayesian belief network model predict reliability of such safety critical digital systems.