Invariant preservation in iterative modeling

Abstract

In a Model-Driven Development project, models are typically built iteratively to better satisfy a set of requirements. Therefore it is crucial to guarantee that one iteration of a model evolution does not hinder the previous version. In this paper, we focus on invariant preservation of behavioral models expressed in Algebraic Petri Nets. The theory developed is applied to a Multi-Level Security File System modeled iteratively. We also discuss how this approach can be applied on Domain-Specific Languages that are translated to Algebraic Petri Nets.