The Evaluation of BH-AllStar Approach in the Industry: A Case Study

The software testing process plays an important role in the software development life cycle. Test data generation is a critical step for efficient software testing. In the industry, safety-critical software tests must be carried out in a way that ensures software safety standards. Combinatorial Testing (CT) is one of the most frequently used method while generating test data in order to meet standards and also to perform efficient tests with higher code coverage rates. However, the test data explosion is a big problem in CT. To overcome this challange, various meta-heuristic methods are developed by using Search Based CT (SBCT) methods. In this study, the usability of a search-based meta-heuristic method (BH-AlIStar) in the industry, which was developed based on the Black Hole Algorithm (BHA), has been investigated. To achieve this goal, we used a more complex safety-critical industrial Software Under Test (SUT) than the one used in the reference study. Then, we evaluated basic BHA and BH-AIIStar on this SUT with respect to code coverage rate, execution time and number of test cases. According to the results, the BH-AllStar provides a 22% increase in code coverage compared to the basic BHA. This case study strengthens the argument that the BH-AllStar can be used in test data generation in the software industry.