An Empirical Study on the Effectiveness of Automated Test Case Generation Techniques

摘要

The advent of automated test case generation has helped to reduce the laborious task of generating test cases manually and is prominent in the software testing field of research and as a result, several techniques have been developed to aid the generation of test cases automatically. However, some major currently used automated test case generation techniques have not been empirically evaluated to ascertain their performances as many assumptions on technique performances are based on theoretical deductions. In this paper, we perform experiment on two major automated test case generation techniques (Concolic test case generation technique and the Combinatorial test case generation technique) and evaluate based on selected metrics (number of test cases generated, complexities of the selected programs, the percentage of test coverage and performance score). The results from the experiment show that the Combinatorial technique performed better than the Concolic technique. Hence, the Combinatorial test case generation technique was found to be more effective than the Concolic test case generation technique based on the selected metrics.