A Moving Average Modeling Approach for Computing Component-Based Software Reliability Growth Trends

Abstract

This paper introduces a moving average reliability growth model to describe the evolution of component-based software. In this model, the reliability of a system is a function of the reliabilities of its constituent components. The moving average provides a trend indicator to depict reliability growth movement within the evolution of a series of component enhancements. The moving average can reduce the effects of bias or measurement error of certain components by rendering a smoothed trend of system reliability growth. The input parameters are the components’ configurations and individual reliability growths. The output is a vector of moving averaged system reliability growths indicating increasing component enhancement. The application of this model can facilitate cost/performance evaluation and support decision making for future software maintenance. More importantly, without introducing excessive computation, the model can be combined with many existing component-based reliability models to compute overall reliability growth.