An adaptive neuro fuzzy model for estimating the reliability of component-based software systems

Although many algorithms and techniques have been developed for estimating the reliability of component-based software systems (CBSSs), much more research is needed. Accurate estimation of the reliability of a CBSS is difficult because it depends on two factors: component reliability and glue code reliability. Moreover, reliability is a real-world phenomenon with many associated real-time problems. Soft computing techniques can help to solve problems whose solutions are uncertain or unpredictable. A number of soft computing approaches for estimating CBSS reliability have been proposed. These techniques learn from the past and capture existing patterns in data. The two basic elements of soft computing are neural networks and fuzzy logic. In this paper, we propose a model for estimating CBSS reliability, known as an adaptive neuro fuzzy inference system (ANFIS), that is based on these two basic elements of soft computing, and we compare its performance with that of a plain FIS (fuzzy inference system) for different data sets.