NHPP-based change-point modeling for software reliability assessment and its application to software development management

We often observe a phenomenon that the stochastic behavior of software failure-occurrence time-interval changes notably due to the change of the testing-environment during testing activities. Testing-time when such phenomenon is observed is called change-point. We discuss a framework for software reliability growth modeling with the change of the change-point for improving the accuracy of software reliability assessment based on software reliability growth models, and develop specific non-homogeneous Poisson process (NHPP) models based on our modeling framework based on typical NHPP models. Also, we investigate the usefulness of our proposed models by conducting goodness-of-fit comparisons with corresponding existing models by using actual data. Further, we discuss the optimal software release problems, which are interesting issues in software development management, based on our proposed models. Finally, we show numerical examples for software reliability assessment based on our models and the derived optimal policies by using actual data.