Evaluating evolutionary testability for structure-oriented testing with software measurements

Test case design is the most important test activity with respect to test quality. For this reason, a large number of testing methods have been developed to assist the tester with the definition of appropriate, error-sensitive test data. Evolutionary testing is a promising approach for automating structure-oriented test case design completely. In many experiments, high coverage degrees were reached using evolutionary testing. However, evolutionary testing is not equally well applicable to different test objects. For example, evolutionary testing of a test object with complex predicates might fail. In order to assess the difficulty of a test object for evolutionary testing, software measures can be used. The knowledge provided by software measurements could lead to a significant increase in efficiency of evolutionary testing. In this paper, we investigate the suitability of structure-based complexity measures for the assessment of whether or not evolutionary testing can be performed successfully for a given test object.