Continuous Integration in Multi-view Modeling: A Model Transformation Pipeline Architecture for Production Systems Engineering

Background. Systems modeling in Production Systems Engineering (PSE) is complex: Multiple views from different disciplines have to be integrated, while semantic differences stemming from various descriptions must be bridged. Aim. This paper proposes the Multi-view Modeling Framework (MvMF) approach and architecture of a model transformation pipeline. The approach aims to ease setup and shorten configuration effort of multi-view modeling operations and support the reusability of modeling environments, like additional view integration. Method. We combine multi-view modeling with principles from distributed, agile workflows, i.e., Git and Continuous Integration. Results. The MvMF provides a light-weight modeling operation environment for AutomationML (AML) models. We show MvMF capabilities and demonstrate the feasibility of MvMF with a demonstrating use case including fundamental model operation features, such as compare and merge. Conclusion. Increasing requirements on the traceability of changes and validation of system designs require improved and extended model transformations and integration mechanisms. The proposed architecture and prototype design represents a first step towards an agile PSE modeling workflow.