Automated model transformation and its validation using AToM³ and AGG

Abstract:

Complex systems are characterized by components that may have to be described using different notations. For the analysis of such a system, our approach is to transform each component (preserving behaviour) into a single common formalism with appropriate analysis methods. Both source and target notations are described by means of meta-modelling whereas the translation is modelled by means of graph transformation. During the transformation process, the intermediate models can be a blend of elements of the source and target notations, but at the end the resulting model should be expressed in the target notation alone. In this work we propose defining also a meta-model for the intermediate process, in such a way that we can apply some validation methods to the transformation. In particular, we show how to prove functional behaviour (confluence and termination) via critical pair analysis and layering conditions, and discuss other desirable properties of the transformation, namely: syntactic consistency and behaviour preservation. The automation of these concepts has been carried out by combining the AToM3 and AGG tools.

Año de publicación:

2004

Keywords:

• Graph transformation
• Multi-Formalism
• Meta-Modelling
• modelling

Fuente:

scopus