Multi-layer syntactical model transformation for model based systems engineering

Systems engineering supports interdisciplinary decision making in design processes for complex systems by allowing engineers from different disciplines to have an integrated view of the target system. Unlike traditional document based systems engineering, model-based systems engineering (MBSE) uses various types of models to facilitate formal communication among the disciplines, which is essential for system level decisions.;In this dissertation, we propose a new model transformation approach to deal with this variety of models, which plays an essential role in MBSE. Up to now, model transformation has primarily been used to support development processes of software systems in the context of model driven engineering (MDE); applying the model transformation to MBSE, which deals with general systems, gives rise to a number of new problems. We indentify and focus on two key problems: instance data integration for virtually evaluating target systems, and syntactical inconsistencies among commercial engineering tools.;In order to address these two issues, we propose multi-layer syntactical model transformation by extending the standard model transformation methods and tools. We intuitively present the key concepts and the practical benefits of the new model transformation. This intuitive description is supported by theoretical discussion based on graph grammar theory. Finally, we demonstrate the new approach by implementing it in the optimization domain, which is a primary analysis domain of industrial engineering.;This dissertation is organized as follows.;Chapter 1 highlights the importance, and the motivations of this research. It also provides the detailed outline of the dissertation. Chapter 2 explains necessary background knowledge on formal modeling and model transformation. In chapter 3, we propose the new model transformation, the key ideas and concepts.;Chapter 4 provides the theoretic foundation of our multi layer model transformation. We find one necessary condition under which the new transformation is viable. The theoretical discussion is based on the mathematical formalism of model transformation, called triple graph grammar (TGG). We prove two key properties of TGG; these properties play important roles in establishing the theoretical foundation of the multi layer model transformation, and implementing it in the following chapters.;The following two chapters demonstrate the proposed approach. For the demonstration, we use meta-modeling frameworks, and model transformation tools which are built upon the Eclipse development environment. In order to practically support the model transformation, we improve some parts of the tools; and we also propose new ways to use the existing functions.;Finally, we end this dissertation with the summary, the contributions, and the future research topics.