Simulation of Software Reliability Growth Model Based on Fault Severity and Imperfect Debugging

Abstract

The existing software reliability growth model (SRGMs) usually assumes that the detected faults can be eliminated well when considering different types of software faults, to simplify the problem. Therefore, given these existing defects, we propose a new non-homogeneous Poisson process (NHPP) SRGM based on considering different fault severity. According to the complexity of the fault, we define the software fault as three levels: Level I is a simple fault, Level II is a general fault, and Level III is a severe fault. In the process of fault detection, the model comprehensively considers the tester’s ability to find problems and the number of remaining issues. In the process of debugging, the problems of imperfection and new fault introduction are considered. Two kinds of real data sets, fault classification and non-classification, were selected and we made simulation for the proposed model and other traditional SRGMs on the PyCharm platform. The experimental results show that the software reliability model considering fault severity has excellent performance of fault fitting and prediction on both types of data sets.