Automated Method for Software Integration Testing Based on UML Behavioral Models

Abstract

Nowadays, testing is often considered more important than the code itself. Therefore, in order to test large and complex systems test automation methods are needed, which help evaluating whether the software is working properly. The main goal of the research is to improve effectiveness of integration testing by creating an automated method based on UML behavioral models. Test input data generation using symbolic execution was applied and it gave full structural coverage, which increased the quality of integration testing. Testing method allowed automating the testing process and increased the effectiveness of tests in comparison with other methods. Experiments showed that 96% of all mutations were successfully detected, and automated test data generation based on symbolic execution increased the detection of mutants by 6-19% in comparison to other model-based testing methods. © Springer-Verlag Berlin Heidelberg 2013.