本研究乃是探討在一個生產/存貨系統中，製造商向單一供應商提出協調多種原物料聯合補貨的策略，以利運用多種原物料產製一種完成品以符合客戶的需求。此原物料供應商與製造商在一次聯合補貨的週期中，除須斟酌雙方原物料的持有與訂購成本外，還須考量第三方物流業者提供運輸數量折扣的優惠。另外，因為製造商的倉儲空間有限，可能因超出限額部分的存貨，需租用成本較高的委外倉儲，而增加委外倉儲成本。在上述的情境下，製造商採用聯合補貨的策略，決定最佳的完成品製造週期、各種原物料之聯合補貨週期及週期內的運送次數，以期求得生產存貨系統中最低的平均總存貨成本。本研究先針對上述情境建構數理模式，進行理論分析，並提出遺傳演算法與啟髮式演算法兩種方法求算最佳解。本研究依據隨機產生的問題作為資料實驗的基礎，顯示兩方法均有良好的求解效率及解答品質。 This study concerns with a production-inventory system in which the single manufacturer adopts the Joint Replenishment Policy (JRP) to order several kinds of raw materials from the single supplier to produce a finished product to meet consumers』 demand. In a joint replenishment cycle、the decision maker needs to take into accounts the following cost terms. First、one should consider the inventory holding and ordering costs of both parties. Also、one should care about the transportation discounts offered by the third party logistic service provider. Furthermore、the additional warehousing costs from leasing extra warehouse space outside will be incurred when the total warehouse space requirement exceeds the limit. The focus of this study is to determine the optimal manufacturing cycle time for the finished product、the optimal joint replenishment cycle time and the optimal number of shipments from the supplier to the manufacturer for each type of raw material so as to minimize the average total costs in this production-inventory system. To solve this problem、we formulate a mathematical model and propose two solution approaches、namely、a genetic algorithm and a search heuristic. Our numerical experiments demonstrate that the proposed solution approaches are effective in both run time and solution quality using randomly generated instances.
