Software Belief Reliability Growth Model Considering Testing Coverage Based on Liu Process

Software has been increasingly important in many critical systems. To measure the growth of software reliability during testing phases, several software reliability growth models (SRGMs) have been investigated. Different from hardwares, softwares' failure mechanisms follow the basic rules of logic, behavior, and psychology rather than physical laws. Consequently, software testing phase embodies lots of epistemic uncertainties, which are not suitable to be modeled by probability theory-based SRGMs. In addition, fuzzy theory-based SRGMs may give counterintuitive results. Confronted with these, we proposed a software belief reliability growth model (SBRGM) under the framework of uncertainty theory which is a new mathematics system to deal with epistemic uncertainties. However, testing coverage, a key factor in software testing, has not been considered in SBRGM. Therefore, a novel SBRGM considering testing coverage, named SBRGM with testing coverage (TCSBRGM), is proposed in this paper. Based on belief reliability theory, several essential belief reliability indexes for software such as belief reliability and belief reliable time are considered. Unknown parameters in the proposed TCSBRGM are estimated. Finally, real data analysis using two real-world software testing data is carried out to compare the performance of our proposed model with some representative SRGMs. The results exhibit that the proposed model gives a better fitting and predictive performance.