Research On Software Reliability Model Based On NHPP

With the rapid development of computer technology,software is widely used in every field of national defense construction and national economy,the scale is expanding and complexity is gradually increasing.At the same time,how to guarantee the quality of the software has become the focus of attention and research.By improving software reliability is one of the ways to guarantee software quality.And the software reliability model can be used to estimate and predict the reliability of the software through statistical method.However,in order to facilitate modeling,many models were proposed based on unrealistic or over-simplified assumptions.It is obviously unreasonable.This paper mainly focuses on the software reliability modeling based on imperfect debugging.At the same time,the influences of test coverage?test effort and the difference between testing and operating environment have been considered.Based on above,the software reliability has been established.The main contents are as follows:Firstly,the influence of the test coverage is added on the basis of considering the rate of the fault introduction and the elimination rate during the process of decreasing faults.A software reliability model is established,which takes into account the test coverage based on the imperfect debugging.two sets of published data are be used to validate the performance of this new model.Secondly,a new software reliability model is proposed based on the software reliability model considering the test coverage and the effect of the test effort.At the same time,the model is validated by two sets of published data to validate the fitting and prediction ability of the model.Thirdly,in the process of reliability modeling,the difference between testing and operating environment is to be considered.A new environmental factor function is used to build the reliability model based on imperfect debugging.This function can describe the fault detection rate of the testing and operating better.Using the existing data sets,it is proved that the model's fitting ability is strong and the prediction accuracy is high.