Design Retrieval by Fuzzy Neurocomputing

The need to automate parts of a design process has generated a variety of software packages, such as computer-aided design and manufacturing (CAD/CAM), computer-aided engineering (CAE) and others, that have become conventional engineering tools. The new generation of modem three-dimensional geometric modelling systems can indeed assist designers during the conceptual or creative phase of the design process. However, the capabilities of these tools can be highly enhanced by incorporating an intelligent design retrieval system that can automatically retrieve an existing design from the database of design solutions to solve the problem. This paper examines the feasibility of utilizing fuzzy associative memory in design retrieval and illustrates the concept by applying it to a sample problem of selecting an appropriate design for a solar heating system. In particular, a technique in which a design can be automatically retrieved based on how well it satisfies the desired functional requirements or input criteria is introduced, and issues regarding the integration of the proposed system with a commercially available CAD package are discussed.