Optimal Test Case Generation for Simulink Models Using Slicing
Zhenying Jiang, Xiao Jun Wu, Zeqian Dong, Ming Mu · 2017
Simulink is widely used for avionics and automotive systems design within model driven approach. For system verification and validation effectively, it is essential to generate test cases for Simulink models which guarantee high coverage of requirements and completeness required by safety-critical systems certification. However, for large-scale Simulink models, there is limited ability of test generation tools because state-space explosions occur in calculations, and structure coverage of test cases cannot meet the standard because of complex dependency relations in the model. To overcome these drawbacks, we propose a model slicing technique for optimal test case generation, which includes a static slicing algorithm to decrease model scale and generate tests for requirements, and a dynamic slicing algorithm to improve the structure coverage of test cases. Furthermore, we evaluate the effectiveness of slicing approach with avionics Simulink models. The experimental shows that complexity of Simulink models can be reduced and high structure coverage can be reached.