In recent decades Model Driven Software Development became a widespread methodology in software engineering. Its purpose is to facilitate the usage of models through the complete development process starting from the early specification down to the implementation and maintenance phases. Its driving idea is to start the development with the definition of high-level engineering models and though well-defined refinement steps that ultimately leads to low-level platform specific representations that can be used to automatically synthetize source code, documentation, configuration or test cases.
Model-Driven techniques are also started to gain acceptance in the automotive industry mainly due to the AUTOSAR standard. The standard envisages a development process, where the applications running on common ECUs (Electronics Control Units) are first defined as a component based model. This allows that in the early stages of development - without an actual execution hardware – the models can be simulated/tested to validate certain behavior of the system under development. Since testing and simulating these models require similar test cases as in case of ordinary SW development. However, constructing complex structural models by hand is cumbersome that would require some level of automation
In the current thesis, I defined a Design Space Exploration test model generation approach for AUTOSAR, where the main goals was to provide user defined like models (and not randomly generated) and. scale up to industrial size problems.