基因演算法於船体结构构件之拓朴最佳化设计应用TheApplicationof.PDF

基因演算法於船體結構構件之拓樸最佳化設計應用 The Application of Genetic Algorithms on Topological Optimization of Ship Structural Element 計畫編號 ：NSC 89-2611-E-006-030 執行期限 ：88/08/01 ~ 89/07/31 主持人 ：楊澤民 國立成功大學造船工程系所副教授 一 、中文摘要 Abstract 本文主要是應用基因演算法 The state-of-art genetic algorithm (GA) (Genetic Algorithm; GA)對T 加強板架 can rapidly search for the approximate (Tee stiffened panel) 、平板加強板架 global optimum under complicated (flat-bar stiffened panel) 及皺褶板 design environment. Meanwhile it can （Corrugate bulkhead ）等常用且最具代 handle the optimization problem involving discrete design variables. 表性之船體結構件進行最佳化設計 ， The main goal of this study is to obtain 使結構在滿足終極破壞限制(ultimate an optimal design of two stiffened panel failure constraints) 與耐用破壞限制 forms, namely tee stiffened and flat-bar (serviceability failure constraints)之所 stiffened panels, which are often used in 有設定限制條件下 ，求得最佳目標函 ship structures. In this search, these 數值中各設計變數之最佳組合 。在分 stiffened panels are investigated by 析過程中考量不同族群大小 、交換機 means of GA approach. The influences of genetic parameters, such as 率 、突變機率等因素對最佳化結果的 population sizes, crossover probability 影響 。文中是以製造成本為目標函 and mutation probability for each design 數 ，其中同時考量材料成本及勞工成 variable will carefully examined in this 本 ，其所得之結果並與連續性線性規 study. The least cost of panel 劃(Sequential Linear Programming; SLP) production, including costs of materials 最佳化結果作比較 。因此以基因演算 and labors, is defined