Title: Application of meta-heuristic algorithms for automated software test data generation

Authors: Yeresime Suresh; Santanu Ku. Rath

Addresses: Department of Computer Science and Engineering, National Institute of Technology, Rourkela, Orissa, India ' Department of Computer Science and Engineering, National Institute of Technology, Rourkela, Orissa, India

Abstract: The quality of a developed software system depends on factors such as reliability, maintainability, scalability etc. To deliver reliable software, it is necessary that it is defect free. In order to achieve this objective, the software system needs to be thoroughly tested. Since exhaustive testing is not possible, it is necessary to identify suitable test data that helps in improving the quality of a software product. Manual generation of test data is laborious and time consuming due to the presence of a huge number of decision nodes in a program. Meta-heuristic algorithms can help in such scenario by generating optimal test data automatically, from a very large search space of candidate solutions. In this paper, an attempt is made to automatically generate test data using five meta-heuristic algorithms. The proposed approach is applied for a case study (bank ATM). Results show that meta-heuristics are efficient in generating optimal test data.

Keywords: affinity; control flow graphs; CFG; fitness; metaheuristics; software testing; test data generation; software development; ATM software; genetic algorithms; particle swarm optimisation; clonal selection algorithm; CSA; binary PSO; BPSO; artificial bee colony; ABC.

DOI: 10.1504/IJCISTUDIES.2015.071179

International Journal of Computational Intelligence Studies, 2015 Vol.4 No.2, pp.113 - 133

Received: 05 Aug 2013
Accepted: 25 Apr 2014
Published online: 16 Aug 2015 *

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