M. Wimmer, P. Langer: A Benchmark for Model Matching Systems: The Heterogeneous Metamodel Case, Vortrag: International Workshop on Comparison and Versioning of Software Models (CVSM 2013), Aachen, February 27, 2013, Proceedings of the International Workshop on Comparison and Versioning of Software Models (CVSM'13) @ SE'13, (2013), pages 6. pdf
Recently, several model matching tools emerged thatcompute the corresponding elements of two differentmodels. Having a closer look at concrete matchingscenarios, one may find different cases such as match-ing different versions of one model, e.g., model evolu-tion, or matching heterogenous models that do nothave a common predecessor. Especially, the lattercase is interesting as a first step when similar but atthe same time potentially heterogenous metamodelshave to be bridged for exchanging models betweendifferent tools.Most model matching tools provide means formatching heterogenous metamodels. In particular,several heuristics are implemented in matching algo-rithms to reason about heterogenous structures andterminologies used in the metamodels to be matched.However, which set of heuristics is appropriate for het-erogeneous metamodel matching and how to weightthem when aggregating the different results to achievecorrespondences with high quality is still an open chal-lenge. To address this issue, we propose a benchmarkconsisting of real world metamodels and manually de-fined expected correspondences that allows to evalu-ate automatically the quality of the output of modelmatching tools.