Change‐points‐based software scheduling

Abstract

AbstractSoftware reliability literature consists of various change‐point‐based software reliability growth models and related release time problems. The primary assumption of the existing models is the existence of change‐point before software release time only. This does not look practical as the testing team becomes more proficient in detecting the faults due to their continuous involvement in software development by the software release time. Hence the fault detection rate in the pre‐ and postrelease phase is not the same. To capture this change in fault detection rate in the pre‐ and postrelease testing phase, we propose a new software reliability modeling framework by considering two change‐points during the software lifecycle; that is, there exists a change‐point before release time and release time as a change‐point. Further, in the last one‐decade software firms have changed their strategy of stop testing the software after release and continue to test even after release to remove the number of faults to provide better user experiences. This phenomenon attracted academicians to develop theoretical as well empirical study on postrelease testing and formulation of related release time problem. In this paper, we propose a software cost model to determine optimal release and testing stop time considering under the assumption of two change‐points as mentioned above. The proposed model is validated on real‐life data set.