Adaptation and evaluation of generic model matching strategies

摘要: Model matching is gaining importance in Model-Driven Engineering (MDE). The goal of model to identify correspondences between the elements two metamodels or models. One main application scenarios derivation transformations from metamodel correspondences. correspondences, turn, offer a potential address other MDE needs. Manually finding labor intensive and error-prone when (meta)models are large. To automate process, research community proposes strategies combining multiple heuristics. A problem that heuristics limited certain representation formalisms instead being reusable. Another difficulty systematically evaluate quality strategies. This work contributes an approach deal with mentioned issues. promote reusability, consists whose loosely coupled given formalism. systematize evaluation, automatically extracts large set modeling test cases repositories, uses megamodels guide strategy execution. We have validated by developing AML domain specific language on top AmmA platform. By using AML, we implemented library demonstrate our goes beyond context, tested ontology as well. At last, contributed three use show applicability (meta)model interesting topics: co-evolution, pivot synchronization