Evolutionary approach to generating test data for data flow test. Issue 4 (1st August 2018)

• Record Type: Journal Article
• Authors: Jiang, Shujuan; Chen, Jieqiong; Zhang, Yanmei; Qian, Junyan; Wang, Rongcun; Xue, Meng
• Is Part Of: IET software. Volume 12:Issue 4(2018)
• Page Start: 318
• Page End: 323
• Publication Date: 2018-08-01
• Subjects: genetic algorithms -- program testing; evolutionary approach -- data flow test -- software testing -- software effort -- branch -- DUPs -- diverse software development tasks -- definition-clear path -- test cases -- test data generation -- authors -- definition–use pairs -- definition–use coverage -- genetic algorithm; Computer software -- Periodicals; Software engineering -- Periodicals; 005.1
• DOI: 10.1049/iet-sen.2018.5197
• Languages: English
• ISSNs: 1751-8806

Abstract

Abstract : Software testing consumes a significant portion of software effort. Program entities such as branch or definition–use pairs (DUPs) are used in diverse software development tasks. In this study, the authors present a novel evolution‐based approach to generating test data for all definition–use coverage. First, the subset of DUPs, which can ensure the coverage adequacy, is computed by a reduction algorithm for the whole DUPs. Then they apply a genetic algorithm to generate test data for the subset of DUPs. Furthermore, the fitness of an individual depends on the matching degree between the traversed path and the definition‐clear path of each target DUP. They also investigate the coverage and the size of test cases of test data generation by applying the authors' approach on 15 widely used subject programs. The experimental results show that their approach can reduce the size of test cases that generated without affecting the coverage rate.