Model matching - processes and beyond

Conceptual models in general, and process models in particular, have been established as a means to design, analyze, and improve information systems [1]. The creation, utilization, and evolution of such models is supported by manifold concepts and techniques that offer, for instance, re-use driven modelling support, harmonization of model variants, model-based system validation, and effective management of model repositories. Many of these techniques share reliance on the identification of correspondences between the entities of different models, also termed model matching. The accuracy and therefore usefulness of techniques supporting the creation, utilization, and evolution of models is highly dependent on the correctness and completeness of the result of model matching. This tutorial takes up recent advances in matching process models in particular and provides information systems engineers (both practitioners and researchers) with a comprehensive overview of concepts and matching techniques in this domain. We first clarify terminology and essential notions of process model matching and elaborate on use cases in which model matching proved to be useful tool. Then, we review similarity measures that span the textual, structural, and behavioral dimension of models and form the basis of matching techniques. Although we focus on the case of matching process models, the tutorial also outlines how other types of conceptual models can be matched with these techniques. The tutorial further includes a discussion of practical considerations for the application of process model matching, based on insights from the Process Model Matching Contest conducted in 2013 [2]. Finally, we elaborate on open research challenges related to the integration of user feedback, evaluation measures, and the use of model documentation in the matching process.