1 Software Reliability Nonlinear Modeling and Its Fuzzy Evaluation

Abstract:- Based on the research of the present developing situation in software reliability models, this paper found that it was difficult to satisfy the hypothesis condition of models in actual projects, and these hypothesis restrict the universality of those models. However, the testing data satisfies the properties of time series. This paper presents a nonlinear model of software reliability based on time series and gives corresponding algorithms. The simulation experiments show the accuracy and efficiency of this new model. The newly proposed model suits the requirement of software engineering better and its parameters can reflect the changing of software reliability. The new model can have more accurate analysis and forecast to software reliability issue without any strict assumption. Furthermore, the fuzzy theory is introduced to evaluate software reliability. Using the fuzzy theory that integrated the fuzziness and randomness, the software reliability evaluation method is more approaching to the real system and the accuracy of software reliability evaluation is promoted greatly