In today’s competitive scenario, it is essential for software developers to perform rigorous testing. It helps them to satisfy the demand for reliable software systems. Various external and internal factors like human expertise, fault dependency, code complexity, dynamic approaches etc. affect the process of fault removal. Hence, along the time-line fault removal rate may change. The point on time-line beyond which rates are altered is termed as the change point. Also in many practical situations, the number of failures experienced may not coincide with the number of faults removed from the system. This ratio is computed by Fault Reduction Factor (FRF). Here, we have proposed testing effort based model considering effort-dependent FRF with and without change point for gauging the failure pattern of a software system. The FRF has been modelled by logistic curve. The developed models have been verified using real-life software fault datasets. Model parameters are estimated and various performance criteria are employed to check the goodness of fit. Later, we have developed a software cost model to determine the optimal Release testing effort that minimizes total expected cost of fault removal during testing phase and operational phase of software life cycle subject to a reliability constraint. A numerical study has been also taken to demonstrate the results. Cost sensitivity analysis has been carried out to identify the crucial cost component and role of each cost component on optimal testing effort and overall cost of development.
Testing effort based software reliability assessment incorporating FRF and change point
