遺傳演算法為人工智慧最佳化演算法之一,具易於程式編碼及不依賴函數的梯度訊息等特點,可被用於求解目標函數較為複雜之最佳化問題,但實際應用時普遍存在局部搜索不佳且易於陷入局部極值的窘境。在不改變傳統遺傳演算法運算邏輯的前提下,本文提出結合粒子群演算法的局部模型觀念,取代傳統遺傳演算法的突變操作以提高局部搜索能力,並引入模擬退火法的狀態轉移概念,以增加跳脫出局部解的機會。本文應用所研提之混合式遺傳演算法於斜張橋鋼索預力最佳化設計問題,並以二座斜張橋為案例進行分析、比較與探討。結果顯示,本文所研提之混合式遺傳演算法不僅可於較少之世代數即接近收斂,且所得斜張橋之結構行為亦優於傳統遺傳演算法者,研究成果可供為斜張鋼索預力最佳化設計之參考。
With the advantages of easy coding and effortless mathematic-compiling, genetic algorithm (GA) is a popular optimization solver of artificial intelligence. However, converging to the local optimum may be a general problem of it. In order to overcome this drawback, a hybrid optimization algorithm integrating local model of particle swarm optimization (PSO) and conventional GA to speed up efficiency as well as increase accuracy for global optimization was proposed to deal with optimum post-tensioning cable force in design of the cable-stayed bridges. Two case studies were performed and discussed. The results obtained shows the proposed method has a better performance than conventional GA and could benefit the engineers in optimual design of the cable-stayed bridges.