Applications of Stochastic Differential Equation and Bayesian Analysis on Software Reliability Growth Models
Software reliability engineering is critical for software industry since it can provide useful information for software developers and testing staff during the testing phase. In recent years, several software reliability growth models (SRGMs) help software industries to develop such software which is error free and reliable. Moreover, most of SRGMs were developed based on the assumption of non homogeneous Poisson process (NHPP) since it can explain the phenomenon of software reliability growth in statistical analysis during debugging phase. However, the issues ofaccurately estimating SRGMs’ confidence intervals and dealing with models’ parameters under insufficient historical data will be a challenge to software managers. Accordingly, in this study, the two theoretical models will be proposed to extend the applications of SRGMs in order to refine the decision quality of software release. Stochastic differential equation and Bayesian decision analysis will be applied to Ohba’s inflection S-shaped model for demonstrating the proposedmodels.
Index Terms - Software reliability; confidence intervals; stochastic differential equations; Bayesian analysis; Non-homogeneous Poisson process