Comprehensive analysis of hybrid nature-inspired algorithms for software reliability analysis. Issue 6 (17th August 2020)

• Record Type: Journal Article
• Authors: Sangeeta, ; Sitender,
• Is Part Of: Journal of statistics & management systems. Volume 23:Issue 6(2020)
• Page Start: 1037
• Page End: 1048
• Publication Date: 2020-08-17
• Subjects: 68-04; Software reliability -- Parameter estimation -- SRGM -- GA -- ABC -- PSO -- DE -- FP -- PSOGSA; Statistics -- Periodicals; Mathematical models -- Periodicals; Mathematical models; Statistics; Periodicals; 519.5
• DOI: 10.1080/09720510.2020.1814498
• Languages: English
• ISSNs: 0972-0510

Abstract

Abstract: Software reliability growth model accuracy could be justified only with the help of their parameter estimation capability. Closer are the estimated parameters to the actual failure dataset higher will be the accuracy of that software reliability model. Inaccurate estimate of the parameters by the software reliability growth model may lead to heavy losses. Authors in this paper recognized nature inspired meta-heuristic algorithm based effective methods for parameter optimization of software reliability models. Both interval domain and time domain datasets have been used for software reliability model parameter estimation process and experiments have been conducted using real project datasets. Results are analyzed using actual number of failures given in real datasets and compared among various existing meta-heuristic algorithms. In this paper, authors also identified the use of hybrid meta-heuristic techniques in comparison to other meta-heuristic algorithms for software reliability assessment and evaluated them based on various performance criteria. Based on the results it is obtained that hybrid algorithms are very much satisfactory in terms of accuracy in parameter estimation as compared to their counterpart and might be used on other software reliability models in-order to assess reliability of a system with higher accuracy.