Towards Intelligent Systems Supporting Conceptual Design

Abstract

This paper deals with an intelligent system which supports the conceptual phase of design. The proposed system makes it possible to extract design knowledge from the early solutions represented in the form of generic design drawings. This knowledge is used to support rapid decision-making throughout the design process. The drawings constitute a mean of human-computer interaction and are internally represented with the use of graph-based structures, which encode both syntactic and semantic aspects of design solutions. The design knowledge stored in graph structures is translated into logic formulas describing design drawings. The presented logic-based reasoning mechanism enables the system to check whether designs satisfy specified requirements and constraints. The proposed approach is illustrated by examples of designing floor layouts, which are used to navigate robots in them and test which layout elements are in the range space of cameras.