A New Idea for Search Based Software Testing: Optimization of Test patterns by Using Local Meta-Heuristic Search

Shojaedini, Seyed Vahab; Delkhoshian, Fahimeh; Khodadoust, Mohsen; Momenian, Mohammad

doi:10.22034/ijiems.2024.429614.1065

A New Idea for Search Based Software Testing: Optimization of Test patterns by Using Local Meta-Heuristic Search

Document Type : Original Article

Authors

Seyed Vahab Shojaedini ¹

Fahimeh Delkhoshian ²

Mohsen Khodadoust ³

Mohammad Momenian ⁴

¹ Department of Electrical Engineering and Information Technology , Iranian Research Organization for Science and Technology (IROST)

² Faculty of Engineering, Alzahra University, Tehran, Iran

³ Department of Electrical, Biomedical and Mechatronics Engineering, Qazvin Branch, Islamic Azad University, Qazvin, Iran.

⁴ Department of Computer Engineering, Faculty of Engineering, Islamic Azad University, E-Campus, Tehran, Iran

10.22034/ijiems.2024.429614.1065

Abstract

Automated software testing is the most important method to ensure from reliability of the software during its performance. The most critical part of this procedure is to generate optimal test patterns. Search based software testing methods, try to generate the test patterns in such way that maximizes the detected errors in parallel with minimizing the time of test. In this paper a new search-based idea is introduced for generating optimized test patterns which is based on local meta-heuristic search. Performance of the proposed method is evaluated in parallel with some state of art algorithms in order to compare their effectiveness in generating optimal test patterns. The above comparisons were performed separately in two fast and slow performance classes. Fast performance class comparisons, show the superiority of the proposed scheme against its alternatives in such way that it improves pattern generation procedure maximally 73 percent and minimally 47 percent compared to alternatives. Such superiorities are also observed in slow performance class, by improving pattern generation procedure maximally 47 percent and minimally 27 percent.

Keywords

Search-based Software Testing

Test Pattern Generation

Local Meta-heuristic Search

Genetic Algorithm

Memetic Algorithm

Opposition Based Learning