Author(s): Dinesh K. Sharma, Deepak Kumar, Shubhra Gautam
With an increasing demand of software systems in today’s society, it is imperative to prepare systems with the utmost reliability. Software reliability is the probability of a system to be free from failure for a given period under given conditions. Software Reliability Growth Models (SRGMs) are used to measure the quality of the software. Many SRGMs assume that software reliability is a one-stage process. However, some researchers consider it to be a twostage process for fault observation and its removal. Further, it is examined in the literature that software fault is imperfectly removed. s debugging may occur in two ways, i.e., imperfect fault removal and error generation. In this paper, we propose two new SRGMs with imperfect fault removal and error generation using learning function. The proposed models are validated on real software data sets and compared with the existing models. Also, the proposed models can be reduced into existing models depending upon the value of the parameters.