本文先利用遺傳法則內選擇、交配與突變三個基本運算子提出一種遺傳演算法,檢討在此遺傳演算法中有與否保留上代族群中幾個最佳解與控制參數之不同給定值對搜尋最後結果之影響,並以幾個數值例來驗證並檢討之。 Warshp程式為一分析貨櫃船扭轉強度的程式,而遺傳演算法(GA)是一種族群搜尋策略的全域最佳化方法。文中以Turbo C++語言在個人電腦上撰寫一界面程式將GA程式與Warshp程式建構為設計分析之整合系統軟體WAGA,最後以一3494TEU貨櫃實船例做測試,並對遺傳演算法內控制參數之不同給定值對搜尋最後結果之影響進行探討,所得的結果證實此整合系統WAGA是可行的。 Based on the principle of selection, crossover and mutation in the traditional genetics, a genetic search algorithm is proposed in this paper, firstly. The influences on searching results of some optimization problems using the present genetic search algorithm with and without elitist model and giving parameters with different values are studied, respectively. Warshp is an analysis software system in dealing with the torsional strength of container ships with large hatch openings. In this paper, an interface program written in C++ language on personal computers for coupling the Warshp software system and an optimization software program, GA, based on the present genetic algorithm, is presented. WAGA is used to solve a structural optimization for a container ship 3494TEU, and the influences of parameters on the optimum design are examined, respectively. The results of the example show that the performance of the integrated system is acceptable.