考量貨櫃運輸成本之聯合補貨問題

Abstract

聯合補貨問題(Joint Replenishmet Problem, JRP)為一因應全球化市場與產品多樣化需求特質所衍生的一個研究課題，藉由產品的合併訂購以降低總成本值，傳統的聯合補貨問題考量的成本項目為(1)主要整備成本、(2)次要整備成本以及(3)存貨持有成本。本研究主要探討在考量運輸成本與固定補貨基本週期的限制下之聯合補貨問題，因此在本研究中除了考量傳統聯合補貨問題具有的成本項目之外，另外亦納入了運輸成本做為成本計算的考量。本研究對於聯合補貨問題建立一個融入上述各項成本項目之數學模式，為了發展一套有效的演算方法以求得聯合補貨之最佳補貨乘數值、最佳補貨排程以及其貨櫃數目，本研究運用基因演算法(Genetic Algorithm, GA)融入區域搜尋的概念設計一個兼具求解效率與求解品質之演算方法，並且對於本研究所建立之數學模式進行求解。最後將其實驗數據結果與LINGO軟體以及共同週期法(Common Cycle Approach, CC)所獲得之數據資料進行比較分析，以實驗數據證實本研究所發展之求解方法為一優良的求解方法。

Researchers are interested in studying the joint replenishment Problem due to its applications in the globalization of markets and products diversification. Also, the practitioners are motivated to reduce the total costs by replenishing products jointly. The costs in the traditional Joint Replenishment Problem are composed of three parts：(1) major setup cost, (2) minor setup cost and (3) inventory holding cost. This study focused on additionally taking into account the transportation cost subject to using a fixed basic period for the planning of replenishment. We formulate a mathematic model by including the above costs in this study. We propose a genetic algorithm (GA) that was equipped with a local search to effectively obtain the best replenishment multiplier value, the best replenishment schedule and the container numbers. Using our random experiments, we compare the proposed GA with LINGO and the Common Cycle Approach, and conclude that the proposed GA is able to efficiently obtain better solutions than the others.