Due to the paradigm shift towards Industry 4.0, the role of software-intensive systems is becoming more and more important. In particular, the trend towards physical components being controlled by software has led to the Internet-of-Things (IoT) and Cyber-Physical-Systems (CPS). As a consequence, companies face highly complex systems that are undergoing a constant change process resulting from shorter innovation cycles and rapidly changing customer needs. It is important that they keep their high-level requirements organized and consistent over multiple revision cycles across the entire life cycle of such a system, i.e., from design over development to implementation and operation.
Modeling is considered a promising technique to better understand the dependencies within such complex systems. By following the Model-Driven Engineering (MDE) paradigm, systems are developed on a higher level of abstraction, and therefore, models are used as an integral part covering requirements, analysis, design, implementation, and verification. Although the term ?model-integrated computing? has been coined almost twenty years ago, it has to be emphasized that the integration of models in the system life cycle is still mainly concerned with forward engineering, i.e., the development of new systems through generative techniques. Much less effort in MDE is spent on the evolutionary aspects of systems changing over time. For tackling this issue, models must no longer be considered as isolated one-shot system prescriptions, but as evolutionary and reusable descriptions of reality.
The research scope of this cumulative habilitation thesis is explicitly addressing this evolutionary aspect by focusing on temporal aspects of models of CPS. It follows a Model-Driven Systems Engineering (MDSE) approach by identifying and integrating appropriate concepts, languages, techniques, and tools for the systematic adoption of models throughout the engineering process. Models are continuously revised, often by considering feedback from other resources, until they are released. However, also the feedback after the release, i.e., from the operation, is reflected in the models. In the first part of this cumulative habilitation thesis, we elaborate on the integration of data from heterogeneous sources in order to provide a homogenized meaningful stack of information from the running system to a higher level of abstraction. In the second part, we cover the evolutionary aspects of engineering artifacts, i.e., models. Thereby, the focus is not only to represent the current state to steer the system but on the representation of the system?s history. In the final part, we provide MDE techniques for analyzing runtime data and extracting descriptive models for reasoning about and validating the operation of systems.