階切策略已被公認為是解模糊最佳化設計問題的基本方法，但是設計者通常不易選擇適當的階切值α。 本文應用本研究群所發展的兩種階切模糊策略，唯一解單切法和修正型雙切法，以求得最適當的改善階切α值。 另外，在結構最佳化的設計問題裡，設計目標函數和限制條件函數很可能屬於隱函數型式，造成求解的困難。 本文應用回應表面法將結構設計的行為響應函數轉換成顯函數型式，結合本研究的唯一解單切與雙階切的模糊解策略，以解模糊結構最佳化設計問題。 進一步以十桿桁架設計例題來表達唯一解單切法和修正型雙切法於結構模糊最佳化設計的應用與價值。 To solve a practical design optimization problem in conjunction with the design of experiment containing fuzzy information, this paper introduces a single level-cut and another double level-cut approach combined in the response surface technology by sequential quadratic programming. A step-by-step algorithm has been presented for a designer easily follows to programming. A ten-bar truss is adopted as the example for verifying the effectiveness of the presenting approach. The results show that the double level-cut approach is the most recommended for the usage in the engineering design optimization with response surface approach.
