A METHOD FOR OPTIMAL DESIGN OF HIGH-TECH PRODUCTS USING CBR AND NEURAL NETWORK

Abstract

Experience plays an important role in solving problems. For that reason, case-based reasoning (CBR) has been utilized in various fields including product design. The experience of expert designers is an important aspect of their capability, helping them to solve problems efficiently and effectively. In chemical applications such as photolithography, the manufacturing process is influenced by various variables including product features, materials, and production conditions. Therefore, the nature of the goods manufactured in such Processes requires engineers to solve design problems based on. past experience rather than theoretical analysis. This paper presents a CBR method for optimal design of high-tech products such as plasma display (PDP) rear panels. We develop an inference engine that can, prior to pilot product testing, reason and predict design results based on past development cases. A design case study using cluster analysis and a radial basis function network to Predict design results of the photo-etching process is presented.