遺傳演算法於配送點選擇之應用

Application of Genetic Algorithms on the Capacitated Plant Location Problem

配送點位置的選擇是在規劃一個製造或服務配送網路時的一個核心問題。由於遺傳演算法對解答空間的要求較少，我們根據決策者的實際需要，建立了一個採用配送點設置及關閉成本、貨物運送成本、以及貨物持有成本這三個重要因素的多因子非線性數學模型。另外，我們針對突變在搜尋早期的重要性以及晚期造成偏差的缺失，提出了一個採用適性搜尋策略的遺傳演算法，並以嚴謹的測試驗證此適性搜尋策略的效益。

The application of genetic algorithms (GA) on the capacitated plant location problem is studied in this paper. Since GA imposes less constraints on the solution space, a more realistic plant location model is constructed with three most important factors: facility setup and closing costs, transportation cost, and inventory holding cost. In addition to constructing the new model, we incorporated an adaptive search strategy in traditional GA to improve its effectiveness. A large number of experiments were performed to demonstrate the effectiveness of this adaptive search strategy.