Multi-Domain Dependency Graphs for Analyzing Cyber-Physical Systems

Cyber-Physical Systems (CPS) combine hardware, software, and physical processes from different engineering fields, and often exist in different product variants and versions. Working across these domains introduces numerous connections and dependencies that make it challenging to analyze systems, identify errors, and track how changes impact the entire product. A major challenge is the lack of a single, unified system model that can represent all system elements, their relationships, and the extra information created during development, such as metadata about authorship, testing results, or change history. Multi-Domain Dependency Graphs (MDG) are introduced as a graph-based modeling approach to address this need. MDG provides a unified system model that includes not only the core technical artifacts from different engineering domains, but also metadata, support for managing product variants and versions, and the ability to capture derived analysis results. The MDG structure uses layers to separate core system elements, variant and version information, metadata, and computed data, making it possible to filter and focus analyses on relevant parts of the system. ... mehrBy supporting integrated feature models and enabling systemwide analyses, such as change impact assessments and clustering, MDG provides a practical foundation for managing CPS in a consistent and scalable manner. Organizing information in layers helps reduce the size of data that needs to be processed and improves the ability to extract insights across domains, making MDG well-suited for advanced engineering tasks and applications involving graph-based AI.