Knowledge and reasoning in design systems

This paper deals with problem-oriented visual languages used for solving design tasks during the conceptual design phase. Visualization of early design solutions by means of visual languages requires an appropriate design knowledge representation in computer systems. The continuation of our method to develop a knowledge-based decision support design system is presented. It consists in connecting a graph-based knowledge representation with a design reasoning method which uses first-order logic formulas. The design diagrams being elements of a visual language are automatically transformed into hierarchical hypergraphs and then information stored in these hypergraph data structures is translated to formulas of the first-order logic forming design knowledge. Modifications of design diagrams cause changes in logic formulas, which can be traced by the designer. The approach is illustrated by examples of designing floor layouts and navigating robots in them.

Research Highlights
► A new methodology for conceptual design supported by a computer system is proposed.
► Visualizations of design solutions are represented by graph-based data structures.
► Design knowledge is stored in hierarchical graphs and logic formulas.
► Reasoning about designs is based on first-order logic formulas.
► The approach is illustrated on examples of designing layouts and robot navigation.