物料需求規劃(material requirements planning, MRP)運用多年，約75%以上的製造業使用物料需求規劃做為主要生產計劃與管制的工具。計畫前置時間(planned lead time, PLT)是現今生產規劃系統一個極重要的參數，它建立生產流程中再製品(work-in-process, WIP)水準，以及存貨控制的準則。較長的前置時間雖然可以使生產更加平滑，針對需求的不確定性去做緩衝。但是，研究指出超過95%以上的前置時間不是真正的生產時間，其中包括等待時間、轉換時間等等，皆遠遠超過實際生產時間，而降低存貨使用的效率，也增加了不必要的生產成本。本研究針對組裝式產業(如汽車、精密製造、電子產品等)為主，並利用混合整數規劃(mixed-integer programming, MIP)，建立一個數學模式，並利用「IBM ILOG CPLEX Optimization Studio」最佳化開發環境，帶入工業界實際數據進行實驗驗證，找出計畫前置時間之最佳值，幫助決策者在運用物料需求規劃系統時制定生產計畫，並能有較佳的運行績效。MRP systems are a common means of developing production plans in discrete parts manufacturing. A key parameter of MRP systems is the planned lead time (PLT) of each part number. MRP systems take the net requirements of a part and offset them backward in time by the PLT to create planned order manufacturing releases. We establish long lead time to smooth the production and to buffer for demand uncertainty. It has already been thoroughly demonstrated in the research that the majority of the lead time, over 95%, is non-productive time. Lead time includes elements like queue time, waiting time, transfer time, etc., all of which are much larger than the actual production time.This paper applies well to assembly industries would include manufacturers of automotive parts, precision casting and electrical equipment. The heart of this paper is a mixed integer programming model for determining optimal PLT. And we do experiments to show that our approach is better than prior approach for setting PLT by using IBM ILOG CPLEX Optimization Studio. To help businesses when use MRP systems to create the more effective production plans.
