Testing-Effort Dependent Software Reliability Assessment Integrating Change Point, Imperfect Debugging and FRF

Abstract

With tremendous growth in the use of software in today’s hi-tech world, it is necessary to have high-quality, reliable software systems. Many SRGM’s have been developed to assess reliability growth. In this paper, an SRGM incorporating Weibull effort function has been investigated. The time point at which the fault detection and removal rates are changed is termed as change point. The testers are unable to perfectly remove a fault on its detection and may add to the number of latent faults. This scenario of imperfect debugging leads to error generation. Here, we propose a testing-effort dependent software reliability growth model incorporating fault reduction factor (FRF), change point and imperfect debugging. The model has been derived under two cases: (i) when FRF is represented by a logistic function and (ii) when FRF is constant. For testing-effort expenditure, we have used Weibull function. The proposed models are tested on two real software fault datasets, and results are compared with perfect debugging scenario. The goodness-of-fit analysis results show the accuracy of proposed models to represent the failure behaviour of the software system.