Abstract
The purpose of this research is to develop a method of using neural network to model machine behaviors, so that we can use trained neural models to 1) help machine parameters setting when a given set of machine performance is desired, and 2) to help machine selection when multiple machines are presented given a set of machine performance or attributes are desired. The research used a two-layer backpropagation network to model molding machines used in the packaging industry. Real data from two I.C. packaging plants are used. The result shows that the method is feasible. Benefits of the approach include: reducing trial experiments, in machine setup for new products, thus reducing waste of materials, labors and times, and finding a intelligent method for equipement selection and parameter setting.