RANKING OF SOFTWARE ENGINEERING METRICS BY FUZZY-BASED MATRIX METHODOLOGY

Abstract

In present day a lot of software’s are being developed day by day. To check reliability of software is a big issue. For this software engineering metrics is formed and ranked by different methods. In present thesis work we proposed a method for ranking of software engineering metrics based on expert’s opinions elicitation and fuzzy-based matrix methodology. The proposed methodology has ability to translate the vague and imprecise data concerned with the problem of ranking of software engineering metrics, and the ambiguity and uncertainty occurring at the time of expert decision making to remove the complexity of formulation of the intention and the control function. The matrices provide themselves to mechanical manipulations and are helpful for evaluating and developing systems functions which match with the purpose of research work. This research work is based on software engineering metrics acknowledged in an earlier study conducted by Lawrence Livermore National Laboratory. A set of ranking criteria were recognized. After that software engineering metrics are ranked in ascending series using experts’ opinion in according to the values of Permanent function on their criteria matrix. The proposed methodology has also been compared with other known methodologies. The use of fuzzy set theory improves the decision-making procedure by considering the vagueness and ambiguity prevalent in real-world system. We also found that the use of triangular fuzzy numbers made data collection, calculation, and interpretation of results easier for experts.