University of Stuttgart

The design of complex cyber-physical systems [1] involves the concurrent development of hardware and software. Furthermore, the current design and development processes for engineering complex systems is characterized by heavy multi-disciplinary coupling across disciplines. The design process of such complex systems involves therefore several domain specialists and follows a systems of systems approach.

The system of systems approach involves decision making on the system topology (i.e. the architectural design decisions) and the system parameters (i.e. the dimensioning of the design components). Since many disciplinary models are used in disciplinary analyses, the consistency between these models in the automated model generation process plays a crucial role for a successful automation of the model generation process and occurs frequently in an iterative design process.

The Evolution of Design Languages

The Evolution of Design Languages

Additionally, for a successful development of a framework which consistently supports the whole product life-cycle needs powerful means to equally well represent, manipulate, simulate and evaluate design and manufacturing knowledge. In order to interactively assist and support, semi-automation or complete automation of such a design and manufacturing development process, a means to represent the design and manufacturing knowledge needs to be developed. The so-called design language workbench for which the University of Stuttgart is held responsible, intends to solve several important issues in this problem setting. Among these are:

    1. Representing engineering knowledge in a readable and digitally processible way by both humans and machines.

    2. Decomposition and structuring of the engineering design knowledge in form of a design language with a vocabulary (i.e. building blocks), rules (i.e. building knowledge) and process knowledge (i.e. building rule sequence).

    3. Allowing the merging, mapping and extension of the knowledge representation in form of design languages by processing mechanisms insuring consistency and correctness

    4. Model generation of all necessary disciplinary engineering analysis models by compilation of the design language into consistent, domain-specific models.

The sought-after engineering design language workbench is based on the representation of both globally generic engineering background knowledge and locally specific engineering product design and manufacturing knowledge in a re-useable engineering ontology. Here in this context, the term ontology is assumed to account for a finite list of interrelated engineering concepts, for which the interdependencies may be expressed as associations and inheritances. For this purpose, the concept and representation format of so-called graph-based design languages on the basis of the Unified Modeling Language (UML) will be used and partially extended.

Since the creation of such an equally globally generic and locally specific knowledge representation involves the cooperation of several specialists, as a consequence several means have to be developed in order to ensure the capability of cooperation of specialists separated in space and time (i.e. the support of concurrent and geographically distributed engineering concepts) and to automatically merge and integrate their partial ontologies into a globally consistent and system-wide accessible and valid re-useable knowledge representation.

Graph Based Design Languages

Graph Based Design Languages


[1] Sztipanovits, J. et al.: Toward a Science of Cyber-Physical System Integration. Proceedings of the IEEE, Volume 100, Number 1, January 2012.