Abstract

Today, model refinement has a key role in software engineering. It is very important to find an approach to add details to the abstract models. This stepwise refinement should continue till the final software is developed. Since this process in most of the cases is boring, long and error prone, it is crucial to find an automatic yet precise approach to perform refinement. In this paper, we present an accurate approach to refine abstract models into more detailed models. To do so, we use graph transformation systems. In the proposed approach, at first, we design the initial abstract model by graph transforming systems. Then, based on the desired meta-model, we design proper rules to generate a more detailed model from the abstract one, automatically. In this approach, both structural and behavioral parts are considered for modeling. At the end, using the formal background of graph transformations we prove the soundness of the approach.

Key words: Model refinement, graph transformation systems.