小領域の特性に着目したGA困難なリブ配置の設計

概要

論文の詳細を見る
The genetic algorithm (GA) is effective for designing the stiffener layout for a plate. In general, however, the coding of the stiffener layout makes the string considerably long, and the schema, namely the effective part of the string, is easily damaged during the genetic operations. This causes the GA-hard problem where the search of the optimum layout frequently fails. A method is proposed to overcome the GA-hardness in the maximum stiffness design of a plate. The plate is divided into several regions to shorten the string, and the stiffener layouts in all regions are optimized concurrently by the aid of the knowledge of the bending and twisting moments in the regions which are expected to be averaged by the optimization. The proposed method is applied to the design of the stiffener layout for a rectangular plate with GA-hard problem, and its effectiveness is demonstrated and discussed.
一般社団法人日本機械学会の論文
1996-07-25