A reliability improvement predictive approach to software testing with Bayesian method

The capability of improving software reliability is one of the main objectives of software testing. However, the previous testing methods did not pay much attention to how to improve software testing strategy based on software reliability improvement. The relationship between software testing and software reliability is very complex and this is mainly due to the complexity of software products and development processes. The software testing strategy with improving reliability on line needs to possess the ability to predict reliability. Model predictive control provides a good framework to improve predictive effect on line. However, one of the main issues in model predictive control is how to estimate the concern parameter. In this case, Bayesian method is used to estimate the concern parameter: reliability. This proposed reliability improvement predictive approach to software testing with Bayesian method can optimize test allocation scheme on line. The case study shows that it is not definitely true for a software testing method that can find more defects than others can get higher reliability. And the case study also shows that the proposed approach can get better result in the sense of improving reliability than random testing.