A New Software Reliability Model for Open Stochastic System Based on NHPP

Among software reliability growth models (SRGMs), the NHPP models perform well in practice. However, the traditional NHPP models still have many problems which are mainly due to the following unreasonable assumptions: 1) fault detection rate subjects to constant or regular change, 2) the testing environment and the final field environment are the same, 3) the testing environment is a completely closed system that does not interact with the outside world. Obviously, these assumptions are not consistent with the reality. Based on the above considerations, we take into account the random interference of the testing system, the uncertainty of the field environment and the openness of the testing system to establish a reliability model which describes the open stochastic system (OSS). Finally, the experimental results show that the proposed model has excellent ability of fitting and predicting.