Efficient run-time co-simulation model switching for holistic analysis of embedded systems

Simulation is widely used for the early validation and evaluation of embedded systems. In the case of complex systems consisting of highly heterogeneous components, a trade-off occurs between the simulation accuracy and time interval under observation. A possible solution for this problem is run-time co-simulation model switching. It enables the test engineer to define the simulation intervals with high accuracy and the intervals of high performance. In this work, we propose a framework for the systematic decision of the component models to be used according to the goals and current status of the simulation. In combination, we further propose a new adaptive approach for run-time co-simulation model switching. The proposed approach is illustrated by the simulation of automotive distributed embedded systems.