Visual Automated Model Transformation

During the design process of dependable computer controlled systems the verification process aiming to assess whether the system fulfils its requirements or needs some re-design is indispensable at each level of abstraction and after each step of system model refinement in order to spare time and resources. When such a refinement step is performed, system characteristics like dependability, timeliness and correctness are assessed. The characteristics are subsequently confronted with requirements, therefore only provenly correct refinement steps are allowed. Several sophisticated mathematical tools exist supporting the assessment of these system parameters (e.g. SPIN model checker [3], PVS theorem prover [6]). However, special mathematical skills and a thorough knowledge of the underlying mathematics are required for their use. Moreover, additional obstacles (such as problems in the consistency and faithfulness of different sort of models) arise due to the lack of an automated transformation and back-annotation method between the mathematical and designer models. The aim of the ongoing research is to provide a general framework for a visual automated model transformation system providing provenly correct and complete, automatically generated transformation algorithms between UML-based system models and different types of mathematical models [7].