二階(S-1,S)庫存政策最佳化模式研究

Abstract

本研究是探討二階存貨系統在（S-1, S）的庫存政策下，如何決定各廣內及配送中心的最佳庫存水準，以滿足各廠內所要求的服務水準。由於訂單前置時間和需求的不確定性，最佳庫存水準及配置不易求得。 本研究提供一啟發式（Heuristic）演算法以解決上述問題。此演算法可概述如下：(1)首先假設配送中心的初始庫存水準為無限大，在滿足大於各廠內給定的服務水準之要求下，求出各廠內倉庫最小之庫存水準。(2)由(1)之結果計算各廠內倉庫之實際服務水準。(3) 利用各廠內倉庫實際服務水準與所要求的服務水準間的差距，計算各廠內訂單在配送中心所能允許的期望等待時間。(4)由允許的期望等待時間與配送中心的期望缺貨數之關係，修正配送中心的庫存數。(5)增加最小訂單期望等待時間的廠內倉庫一個庫存量，利用上述的方法重新修正配送中心的庫存水準，若各廠內倉庫所增加的庫存總量較配送中心減少的庫存量少，則重複(2)-(5)，直到系統總庫存量無法減少為止。最後，本研究以一案例說明此演算法。

This study deal with the determination of stocking levels in a two-echelon inventory system with (S-1-S) policy. The system consists of a central warehouse that supplies spare parts to several local factories. Each factory faces random demand. The lead times to ship a spare part from spare part manufactuting center to central warehouse and from central warehouse to local factories, are also random variables. In this study, a heuristic algorithm is proposed that detremine the stocking levels for each factory and central warehouse to meet the required service levels for each factory. The algorithm can be summarized as follows :(1) Suppose the initial stock level in central warehouse to be infinite. Then calculate the minimum stock levels in each factory to meet the required service level; (2)Using the results from (1), we can determine the real service level in ech factory; (3) By the difference of the required and the real service levels, we can determine the expected acceptable order delay time in central warehouse for ech factory; (4) Use the relation of expected acceptable order delay time and expect numbers of backorder in each factory, we can determine the revised stock level in central warehouse; (5) Increase one unit to the stock of the factory that has minimum expected acceptable order delay time among the factories; (6) Repeat the steps from (2) to (5) then to modify the stock level in central warehouse. Stop when there are no decrease in the system. Finally a case is illustrated to explain the heuristic algorithm.