Model-based testing for execution algorithms in the simulation of cyber-physical systems

The understanding of simulation semantics of a hybrid system is a challenge for computational engineers as it requires expertise in computer science, engineering, numerical methods, and mathematics at once. The testing methods for the execution of a simulation are being researched but not yet applied on the industrial level. Consequently, the semantics of the simulation becomes a critical artifact in the system development process. Embracing it from multiple design perspectives is going to create positive implications at the model, software, and hardware level. Hence, a set of self-testing algorithms for the simulation execution are proposed. Deploying them requires an inclusion of an additional testing dimension in the typical V diagram. Thus, a systematic methodology is conceptualized. A prototypical implementation takes advantage of the Simulink® (1) simulation loop and (2) simulation state retrieval during the system model execution. These two artifacts allow to control and monitor the execution of a simulation. As a consequence, the semantics is systematically considered, its correctness is tested, and the numerical approximation is examined. A case study of a Cyber-Physical System illustrates conceptual and methodological aspects of the proposed algorithms.