Modeling design knowledge on structure

Abstract

This paper is on the modeling of design knowledge. It introduces the concept of design graph grammars, which is an advancement of classical graph grammar approaches. Design graph grammars, as proposed here, provide an efficient concept to create and to manipulate structure models of technical systems. This is interesting from two points of view. Firstly, within many existing tools for design support structural dependencies are represented and processed in a proprietary way. Here, design graph grammars possess the flexibility to model even very specific kinds of domain knowledge while still providing a broadly understood semantics. Secondly, structure models are close to the mental model level of human designers. Design graph grammars concentrate on this level, excluding involved underlying behavioral aspects. This may entail the risk of an oversimplification, but gives design graph grammars the potential to be used within the creative parts of the human design process.