Some Software Reliability Models. Approximation and Modeling Aspects

      Аннотация: Within the hierarchical models in the procedure for quantifying the quality of software products, an important role is played by the so-called computational method based on the theoretical and empirical dependencies, statistical data accumulated during tests, exploitation and the accompaniment of the program product. An important measure of reliability assessment (completeness, accuracy and consistency) is the so-called metric (asymptotic metrics). There are two main approaches to testing: structured and functional. We will only note that, depending on the test data selected and the expected results, testing is divided into: deterministic testing and stochastic testing. In this book, we pay particular attention to both deterministic approaches and probability models for debugging theories. Some of the models have been tested with real-world data. A Hausdorff metric was chosen to evaluate the test data which are fitted to the sigmoid models proposed in this book. Some of the existing cumulative probability distributions are considered in the light of modern debugging and test theories. We propose a software modules within the programming environment CAS Mathematica for the analysis.

Ключевые слова: Hausdorff Distance, Software Reliability Models, confidence bounds, cumulative number of failures, debugging theory, uniform distance, Goel model, Goel-Okumoto model, deterministic curve model, discrete Gompertz, new Gompertz-Makeham, transmuted deterministic, Yamada-exponential, Logistic-exponential, Yamada-Rayleigh, generalized inverted exponential, transmuted inverse exponential, transmuted log-logistic, Kumaraswamy-Dagum log-logistic, Kumaraswamy quasi Lindley