With the advent of the Model Driven Architecture, models are replacing code as the major artifact in software development. A critical success factor for this is the possibility to derive models from each other in terms of transformations. Existing approaches, such as the forthcoming QVT-standard, will provide a proper foundation for transforming models on a fine-grained level. They, however, do not provide appropriate abstraction mechanisms for different integration scenarios, such as integrating models representing cross-cutting concerns or integrating models even from different domains. This paper proposes so called mega operations representing an abstraction mechanism which allow to specify model integration at the metalevel, thus forming the prerequisite to automatically derive a set of directives carrying out the actual integration at the model level. To cope with different integration scenarios, tight as well as loose integration of models is supported on top of a QVT-like language. 1.

Abstract. In state-of-the-art of MDE platforms semantic technologies such as ontologies are rarely used. Our aim is to understand the role of ontologies in supporting model-driven engineering, in particular MDE platforms. MDE platforms may benefit from semantic technologies in formal model semantics and automated reasoning on different levels of the metamodelling architecture. We present an ontology-aware MDE platform architecture and outline some application scenarios where ontologies and automatic reasoning may bring benefit to such platforms. Additionally, an example of using ontologies for verification checks of mapping models in the course of metamodel composition is illustrated.

Conceptual modelling is one of the core disciplines within business informatics. Often, different metamodels have to be integrated to provide project specific or domain specific modelling solutions. This is a challenging task due to the heterogeneity of the modelling languages being assembled. This paper presents an approach for integrating heterogeneous modelling languages in the domain of enterprise modelling. It extends the Enterprise Model Integration approach (EMI) by introducing two key concepts: mappings and rules for metamodel integration. For both mappings and integration rules language definitions are provided. Furthermore, a catalogue of generic integration rules is introduced to provide a set of solutions for recurring metamodel integration problems. The applicability of the presented approach is shown using the case of integrating the Business Process Modelling Notation (BPMN) metamodel with the metamodel for organizational modelling of the ADONIS BPMS method.