Improved novel bat algorithm for test case prioritization and minimization

• Published in: Soft computing (Berlin, Germany) Vol. 26; no. 22; pp. 12393 - 12419
• Main Authors: Bajaj, Anu, Sangwan, Om Prakash, Abraham, Ajith
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.11.2022
• Subjects: Application of Soft Computing; Artificial Intelligence; Computational Intelligence; Control; Engineering; Mathematical Logic and Foundations; Mechatronics; Robotics; Test case minimization; Regression testing; Search-based software testing; Bat algorithm; Test case prioritization; Nature-inspired algorithms
• ISSN: 1432-7643; 1433-7479
• DOI: 10.1007/s00500-022-07121-9

Regression testing is essential for continuous integration and continuous development. It is needed to ensure that the modifications have not produced any errors or faults, thereby maintaining the quality and reliability of the software. The testers usually avoid exhaustive retesting because it requires lots of effort and time. The test case prioritization and minimization solve the issue by scheduling the critical test cases and removing redundant ones. Optimization techniques help by improving the efficiency of these techniques while utilizing limited resources. This paper proposed an enhanced discrete novel bat algorithm for the test case prioritization. The algorithm is modified in two ways. First, we have proposed a fix-up mechanism for the discrete combinatorial problem, which conducts the perturbation in the population using the asexual reproduction algorithm. Second, the novel bat algorithm is improved, where the bats hunt in different habitats with quantum behavior using Gaussian distribution and search in the limited habitat with Doppler effect. In addition, we have embedded the test case minimization procedure in the algorithm for redundancy reduction. The experimental results are empirically analyzed using different testing criteria, i.e., fault and statement coverage on three subject programs from the software infrastructure repository. Consequently, test selection percentage, coverage loss, fault detection loss, and cost reduction percentages are deduced for the test case minimization at program and version levels. Empirical results and statistical comparisons with the random search, bat algorithm, novel bat algorithm, birds swarm algorithm, whale optimization algorithm, and genetic algorithm show the outperformance of the proposed algorithm.