The Research And Implementation Of Software Reliability Model Based On NHPP

In the software life cycle, software reliability engineering is a means of quality assurance. It is a practical science for software quality management and control. In modern society, there may be no artificial materials that can be more ubiquitous than software. Software has become a key part of society in many aspects, for example, home appliances, communications, automotive, aircraft, shopping, network teaching and personal entertainment and so on. People hope that the software quality can make improvements and breakthroughs, so that the software can reach the reliability standard, in order to avoid the economic losses caused by software failures. To ensure that the software system is executed properly and effectively is an important task in the field of software engineering, especially for the manned space system and the nuclear safety control system.In this thesis, the major knowledge system of software reliability engineering is studied, and it makes a deep research on the software reliability model.Firstly, this thesis describes the history of software reliability research, and the current challenges and problems are discussed. It describes the related concepts and theories, and the existing software reliability models have been made the corresponding classification and summary.Secondly, according to the problem of behavior description of fault detection rate, this thesis considers that the inherent fault detection rate is a decreasing function with the increase of time, and the learning ability of the testers is an increasing function over time. This thesis combines the two factors to describe the fault detection rate and proposes the improved software reliability model. Considering the impact of testing effort for the reliability, the model is further amended to improve the prediction accuracy of the model. On the other hand, considering the impact of test coverage for software testing adequacy and effectiveness, this thesis uses the generalized logistic function to describe the change of test coverage rate, and it uses the testing coverage function to derive a new software reliability growth model and fault model. Experiments show that the models have a good adaptability.Finally, this thesis develops a software reliability assessment tool, which makes the evaluation process can be automated. The tool implements the failure data entry, data analysis, parameter estimation, model selection and reliability index calculation. This tool provides a lot of convenience for the software reliability evaluation and prediction, and it can achieve some automation and improve the efficiency of the work.