A Model Orchestra in Digital Twins: A Model-Driven Approach to Integration and Orchestration

This thesis explores the challenges of developing Digital Twins (DTs), which are virtual models of real-world systems used for monitoring, optimization, and lifecycle management. Creating DTs is complex due to the need to integrate and manage different components, especially simulation models (SiMs).

The research focuses on improving how virtual entities, data, and services are handled in DTs. It identifies gaps in current practices, assesses existing tools, and proposes model-driven solutions to make integration and orchestration easier.

This thesis presents studies on practitioners’ practices and tool support for developing digital twins, focusing on integration and orchestration activities in the development process. Based on these investigations, it introduces three main solutions. First, it proposes an approach for generating standardized interfaces, which supports the functional integration of heterogeneous SiMs by unifying how components interact. Second, it presents a method for documenting and contextualizing SiMs using a metamodel, which enables semantic interoperability and facilitates their reuse. Third, it develops a domain-specific language called LOTTS that orchestrates DT services and supports model-to-code transformations, allowing DT systems to be executed automatically and modularly.

These solutions are supported by real-world studies, prototypes, and case studies. Overall, the work shows how model-driven engineering (MDE) can simplify and improve the development of digital twins, making the process more modular, reusable, and scalable.