Ensuring software reliability is a challenging task in software development. Software reliability is the probability of software to provide its intended functionality over a specified time. A couple of testing procedures during the phases of software development provides the data to approximate software reliability. This approximation guides the development team to plan necessary corrective actions. A variety of Software Reliability Growth Models (SRGMs) are in use to predict software reliability. A common task for every SRGM is to calculate reliability growth models attributes as a part of reliability estimation. Optimal calculation of such attributes is influenced by the relationships among the parameters of an SRGM. Therefore parametric SRGMs rely on parameter estimation techniques. The present paper has undertaken the study of existing parameter estimation techniques with the main goal of understanding the pros and cons of each technique in order to design a better technique of parameter estimation for SRGM’s in use. A critical review of existing techniques of parameter techniques is given in this paper detailing the categories, approaches, problems relating to the techniques. One of the most successful swam intelligence techniques named Gray Wolf Optimization (GWO) along with its variants are applied to estimate the parameters of SRGMs. The results of this application along with the comparative analysis are given. The variants of GWO played a significant role in parameter estimation of the SRGMs considered for the experiments. An attempt is made to propose new ways of parameter estimation to achieve optimization.
The Use of Original and Hybrid Grey Wolf Optimizer in Estimating the Parameters of Software Reliability Growth Models
