Authors: D. Ishihara, M.J. Jeong, S. Yoshimura, G. Yagawa and Y.J. Kim
Affilation: University of Tokyo, Japan
Pages: 157 - 160
Keywords: computer-aided optimization, finite element method, design window satisfactory design, continuous evolutionary algorithm, clustering, micro electrostatic actuator
For satisfactory design of micromachines, this paper describes a search method for a multi-dimensional design window (DW), which is defined as an area of satisfactory solutions in a feasible design parameter space. The method consists of the following steps: (1) direct search for satisfactory design solutions using continuous evolutionary algorithm (CEA, a kind of EA), and (2) identification of DWs from the solutions using a modified K-means clustering algorithim. The obtained DWs are expected to give engineers useful information to understand a design problem, and, further, give them new ideas of design. To demonstrate practical performance of the proposed method, it was applied to a shape design of micro electrostatic actuator for optical memory.