Abstract
This paper presents the design method of materials composed of some ingredients by using neural network. In a field of synthetic material design, it is possible to find the elastic properties of synthetic materials from distribution ratio of ingredients using Finite Element Method. But in practice it is largely wanted to find distribution ratio of ingredients from the properties of synthetic materials. Up to now, a trial-and-error technique to do that have been used. Therefore it is required a great deal of labor to find distribution ratio of ingredients. In this paper, the neural network system for material design, that if only the properties of synthetic material are input to the system, distribution ratio of ingredients can be output automatically, is proposed. From the numerical examples of simple material design, it is shown that the proposed method is much available.
