本研究結合基因演算法及模擬退火法兩種不同的最佳化方法，再配合ANSYS有限元素分析軟體，以尋求結構最佳化之設計。基因演算法的優點為採用全域性隨機搜尋的方式選取設計參數值，而在最佳化的過程中僅針對適存函數作處理，如此便可避免複雜的非線性問題，接著以模擬退火法判定解被接受的機率，並有效率地跨過區域最佳解，避免最佳解淪為區域最佳解。最佳化的過程中，吾人以結構最佳化問題之目標函數及限制條件訂定適存函數，經由ANSYS有限元素分析軟體求得適存函數所需之值，而後比較各設計參數之適存度，適存度較高之設計參數經過複製、交配及突變等過程，產生更好的設計參數，再以模擬退火法判斷解被接受之機率，尋求全域最佳值。 本研究將以結構輕量化設計及提高結構之第一模態自然振動頻率為目的，並求得各結構在滿足限制條件時之最佳設計值。數值分析的結果可證明結合基因演算法及模擬退火法於結構最佳化設計可以有效地得到全域最佳解。 In order to obtain the optimum design of a structure, a hybrid method which combined Simulated Annealing and Genetic Algorithm will be adopted in this study. The advantage of Genetic Algorithms is that it has multi-point search strategy instead of one-point search to find the global optimum in a space and the advantage of Simulated Annealing is that it can makes the program have the capability to jump over the local optimum and obtain the global optimum. Therefore, the Genetic Algorithm and Simulated Annealing can be combined together to improve the process efficiency. The new design can be obtained by three basic operators: reproduction, crossover, and mutation. The fitness function in Genetic Algorithm will be developed by finite element software ANSYS program, which is used many aerospace industry. In this study, the minimum weight design and maximum the first natural frequency of different structural problems will be discussed in numerical examples. We hope the results of this study can provide a useful hybrid method for the design of structure.
