Ichiro Nagasaka

Application of evolutionary programming to shape design

Designing the shapes of products is the primary activity of the design process. In this paper, we propose a new model representing free form shape features with the aim of making the system capable of holding and manipulating the shape features after synthesis in order to support the designer early in the design process. The key concept of this study is the shape feature generating process model (SFGP model) which consists of numerous sets of rules using the classifier system (CS). Finally, a computer program was developed to evaluate the model by combining two existing shapes to verify that the shape features of each shape are preserved in the combined shape. It was demonstrated that the model can produce a variety of combined shapes with original, often exaggerated, features.

Adaptive-growth-type 3D representation for configuration design

In this paper, the design of geometrical shapes of function carriers and their layout in given space is called configuration design. The constraint satisfaction problem in configuration design may be difficult to solve due to the lack of tight constraints and the countless combinations of the layout; a diversity of solutions that satisfy the constraints should be allowed. Therefore, to allow such diversity, we directed our attention to developmental processes in biology and proposed an adaptive-growth-type 3D representation based on evolutionary algorithms. Here, the adaptive-growth-type means the shape expressed in the process, which develops through interaction with an outside environment, like shape generation of a living organism in the natural world. The usefulness of the representation was verified by applying it to the component layout problem in the early stage of satellite design.