ORIGINAL ARTICLE Fuzzy based ranking of software reliability measures Anitha Senathi 1 • Gopika Vinod 2 • T. V. Santosh 2 • Dipti Jadhav 1 Received: 9 January 2015 / Revised: 13 April 2015 Ó The Society for Reliability Engineering, Quality and Operations Management (SREQOM), India and The Division of Operation and Maintenance, Lulea University of Technology, Sweden 2015 Abstract This research paper summarizes the results of an implementation of fuzzy multilevel methodology to rank software reliability measures. Of late, computer based systems are used largely for monitoring, protecting and to control safety critical systems like nuclear power plants, Aircraft etc. Reliability is an important factor for assessing the performance of such safety critical digital systems. The characteristics of such digital safety critical systems are explicitly or implicitly reflected by software engineering measures. Therefore, with the help of such measures, models can be built to predict the reliability of software applications that run on safety critical systems. It is not necessary that every software engineering measures con- tribute to predict the reliability, hence they need to be ranked based on their influence on reliability. Since suffi- cient practical data is not available in literature, expert opinion on the selected software engineering measures contributing to reliability based on criterion has been sought. These expert ratings are aggregated and ranked using Chen’s fuzzy logic based ranking method. As the data involved with this kind of problems are inherently imprecise and inexact, application of fuzzy set theory is very suitable for such situations. The top-ranked software engineering measures can be later used to develop a model to predict reliability of safety critical digital systems. Keywords Software reliability  Ranking software engineering measures  Fuzzy aggregation 1 Introduction Software-based digital systems are progressively replacing analog systems in safety-critical systems like nuclear power plants. The modern software-based digital systems consists of hardware and software that controls the hard- ware. Hence reliability assessment for digital systems must consider hardware failure, software failure and effects of hardware and software failures on each other. Incase of hardware reliability evaluation it is assumed that the device design is fundamentally correct. It is also assumed that failures occur due to random variability in the physical stresses experienced by the device and by variability in the devices’ ability to withstand these stresses. However, the reliability of hardware devices can be arrived based on the failure rates under different stress scenarios. Unlike hard- ware, the failures of software applications are mainly due to fundamental design errors. There is no universally ac- cepted qualitative methods for assuring that safety critical software is free of defects to a desired reliability level (NUREG/GR 2000). Software Reliability is defined as the probability that software will not cause a failure of a system for a specified time under specified conditions (ISGOSET 1990). There are many models for predicting reliability of a system, one of them being the ‘‘Early Prediction Model’’ which uses software engineering measures for reliability prediction (Ramamoorthy and Bastani 1982; Lyu Micheal 1996). A measure provides a quantitative indication of the extent, amount, dimension, capacity or size of some attribute of a & Anitha Senathi anithasenathi@gmail.com 1 R.A.I.T, Mumbai University, Mumbai, India 2 Reactor Safety Division, Bhabha Atomic Research Centre, Mumbai 400085, India 123 Int J Syst Assur Eng Manag DOI 10.1007/s13198-015-0359-1