The failure rates of group of software reliability models are used to study the program failure rate per fault at the failure intervals. In these models, it is assumed that failure rate is a discrete function with discontinuities at failure times. It is also assumed that the failure rate of failure distribution is same during the fault detection process which in turn gives the same probability distribution. But the failure distribution can be affected by different factors, such as the software running environment, testing strategy, and the resource. Generally, the running environment may not be homogeneous and may change with the human learning process. Hence, the change point models are of interest in modeling the fault detection process. This paper describes failure rate software reliability model with a change point and also taking into imperfect debugging. The parameters of the model are estimated using maximum likelihood method. A simulation study and a study using real data are used to verify the performance of the model.

software reliability, fault detection process, fault correction process, imperfect debugging, change point, maximum likelihood estimation.