A General NHPP Software Reliability Growth Model with Fault Removal Efficiency
Hongwei Liu, Xiaozong Yang, Feng Qu, Yanjun Shu · Iranian journal of electrical and computer engineering · 2005
This paper presents a general software reliability growth model (SRGM) based on nonhomogeneous Poisson process (NHPP). Although many research have been devoted to unify some NHPP models, most of them either only consider imperfect debugging, learning phenomenon, or take fault removal efficiency as a constant. Consideration of the variation of fault removal efficiency during debugging period in the exiting models is limited. The general model in this paper is the first unified scheme of some NHPP models which take fault removal efficiency as a function of debugging time. Fault detection rate (FDR) is usually used to measure the effectiveness of fault detection of test techniques and test cases. Most researchers assume a constant FDR in deriving their SRGMs. Because of learning process of testers, some researchers take FDR as increasing functions over testing period. Some literature take FDR as decreasing functions because failures removed first have higher detected rate. A bell-shaped FDR function is proposed which integrates both learning phenomenon and inherent FDR. As a special case of the general SRGM, a NHPP SRGM called PBbell-SRGM is put forward which integrates the proposed FDR function and fault removal efficiency. PBbell-SRGM is evaluated using a set of software failure data. The results show that PBbell- SRGM fits the given failure data better than some selected NHPP models.