Application of Adaptive Resonance Networks for Conceptual Design

Abstract: Conceptual design 15 the most important and creative phase of the design process which demands the knowledge, intuition, and imagination of a designer. The Success of the designer in this phase of design depends upon his ability to recall a family of previous design problems which are similar to a current problem and apply the previously discovered solutions. In this paper the application of Adaptive Resonance Theory networks for the creation of families of design problems. characterized by their functional requirements, is examined. A comparison of this method is made with k-mean clustering algorithm. It is found that the Adaptive Resonance Theory networks provide the means for the dynamic creation of families of design problems which cannot be obtained by using conventional clustering algorithms.