A More Practical Software Reliability Growth Model

J. Zhan, X. Zhou (PRC), J. Wang (USA), and H. Li (PRC)


Software reliability growth model, non homogenous Poisson process, distrustable factors.


Up to now, many software reliability growth models (SRGM) have been proposed based on Goel-Okumoto non homogeneous Poisson Process (NHPP) model. Generally, it is assumed that the initial number of faults is a constant at the beginning of software testing. However, pushed by software delivery deadline, the phase of software testing often interleaves with coding in reality. So, the initial number of faults to be detected varies with faults introduced by programmers when testing is ongoing, programmers’ skills and other factors are called distrustable factors. In this paper, we improve the G-O NHPP model by taking the faults introduced by programmers during testing into account. Our study shows that the number of remaining defaults conforms to bio exponential distribution. A numerical example indicates that the new model has a better prediction capability than the G-O NHPP model.

Important Links:

Go Back