Selection of Optimal Software Reliability Growth Models: A Fuzzy DEA Ranking Approach

Over the last 40 years, many software reliability growth models (SRGMs) have been proposed to estimate the reliability measures such as remaining number of faults, software failure rate, software reliability, and release time of software. Selection of an optimal SRGM for use in specific case has been an area of interest for the researchers. Techniques available in the software reliability literature can’t be used with high confidence as they do not provide complete picture about the best suitability of the SRGM for a given real date set. In this paper, we have developed a ranking method to rank SRGM based on fuzzy data envelopment analysis (DEA) approach and then applied it for ranking of SRGMs. The first approach to rank these SRGMs is converting the model parameters set given into linear programming problem by extending CCR model to fuzzy DEA model based on credibility measure level. Since the ranking method involves a fuzzy function, a fuzzy simulation is designed and embedded into genetic algorithm (GA) to establish an algorithm. Finally, numerical example is given to demonstrate the applicability of the proposed fuzzy DEA approach-based ranking method on a real data set.