Title: A state-of-the-art neuro-swarm approach for prediction of software reliability

Authors: Ajit Kumar Behera; C. Sanjeev Kumar Dash; Mrutyunjaya Panda; Satchidananda Dehuri; Rajib Mall

Addresses: Department of Computer Science and Application, Utkal University, Vani Vihar, Bhubaneswar, Odisha, India ' Department of Computer Science and Engineering, Silicon Institute of Technology, Patia, Bhubaneswar, Odisha, India ' Department of Computer Science and Application, Utkal University, Vani Vihar, Bhubaneswar, Odisha, India ' Department of Information and Communication Technology, Fakir Mohan University, Vyasa Vihar, Balasore, Odisha, India ' Department of Computer Science and Engineering, Indian Institute of Technology, Kharagpur, India

Abstract: Software reliability is one of the foremost factors to assess the quality of software. It is evident from the past research that no single model has been developed in the arena of software reliability research to predict the reliability of software. Therefore, lots of attempt is continuously made from different corners of diversity to make a generic and widely acceptable model. In this paper, we propose a neuro-swarm software reliability model by combining the best attributes of functional link artificial neural network (FLANN) and particle swarm optimisation (PSO). FLANNs have been successfully employed to solve nonlinear regression and time series problems; however, its application in software reliability is rare. This intensive work elucidates the feasibility of the use of FLANNs to predict software reliability. PSO is used to tune the parameters of FLANN during the development of the model. The extensive experimental study on a few benchmark software reliability datasets reveals that the PSO-FLANN results is better than models like BPNN, DENFIS, NEBPNN, and canonical FLANN. Hence, the proposed model may be a suitable and promising alternative for predicting software reliability.

Keywords: software reliability; functional link artificial neural network; FLANN; particle swarm optimisation; PSO; normalised root mean square error.

DOI: 10.1504/IJAIP.2021.119020

International Journal of Advanced Intelligence Paradigms, 2021 Vol.20 No.3/4, pp.296 - 322

Received: 27 Jun 2017
Accepted: 10 Mar 2018
Published online: 18 Nov 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article