季節性商品單期兩次訂購動態規劃模型之分析與研究

A Two-Stage Ordering Dynamic Programming Model for Seasonal Product

謝志峰(Chih-Feng Hsieh)

指導教授 ： 蔣明晃 ； 郭瑞祥

國立臺灣大學/管理學院/商學研究所/碩士(2004年)

https://doi.org/10.6342/NTU.2004.00056

摘要

在過去，面對單期季節性商品未知的需求，零售商只能在銷售季節前，粗略評估並一次訂購一整個銷售季節預計可銷售的數量。但是隨著彈性生產、資訊透明化、上下游間的協調整合機制，使得供應鏈快速回應的能力得以加強，打破了過去的限制。在短暫的產品生命週期中，零售商相較於過去，可以在銷售期開始後增加了一次甚至是多次的訂貨機會，並可於銷售初期觀察到實際銷售量後，在更正確掌握需求的狀況下決定後期的訂購量，以提高或保持零售商原有的服務水準並降低缺貨或存貨過多所造成的成本，進而達到利潤的提昇。
本論文有鑑於再訂購策略在數量模型上研究之不足，嘗試建立一個季節性商品單期兩次訂購動態規劃模型，以深入探討零售商在期望利潤極大化的目標下，再訂購策略所得到的利益；本研究並以均等分配及常態分配為例，提出搜尋期初與期中訂購點訂貨量最適解的方法；最後利用實驗設計的方式，探討產品銷售價格、訂貨成本、殘值、變異數、前後半期需求相關係數等因子，對再訂購策略下所帶來的利潤及單次訂購策略利潤差異之影響，以進一步了解再訂購策略所帶來的效益，及相對有效的適用情境。
透過模型之操作與比較，可得知再訂購策略的確對於零售商之期望利潤，有莫大的助益。尤其當前後半期相關性越高，下半期的預測需求可以利用銷售初期的需求資料，作為參考的依據而予以更精確的修正，會使再訂購策略效果更為顯著。

關鍵字

再訂購 ； 季節性商品

並列摘要

In the past, retailers, facing uncertain demands of seasonal product, can only roughly estimate the demand for the entire season, and place an order at once shot before the season begins. With increasing production flexibility, information transparence, and between coordination mechanism, supply chain members can largely improve the capability of quick response. Therefore, retailers can make more than one order in relatively short sales season. Once observing actual sales volume in the initial stage of selling season, retailers can place the second order based on more accurate estimation of demand. That will improve retailer's service level, reduce understock cost and inventory cost, and eventually raise the profit.
Our research first formulates a two-stage ordering dynamic programming model for a seasonal product to catch the interrelationship between two stage's demands in one season. The objective of our model is to maximize the expected profit of retailers. Also, special distributions such as uniform distribution and normal distribution are used to illustrate a searching heuristic, which can find a near- optimal solution. At last, experimental design is performed to find out how the product price, ordering cost, salvage value, demand's variance, and demand correlation between two stages influence the retailer's expected profit and the difference of profit between single order and re-order policy.
Through the analysis of our model, re-order strategy can greatly increase the retailer's expected profit. Especially, as the demand information can be obtained during the early stage and the demand correlation between the two stages is higher, the effect of re-order policy is more significant.

並列關鍵字

seasonal product ； re-order

參考文獻

2.Azoury, Katy S. & Bruce L. Miller, （1984）“A Comparison of the Optimal Ordering Levels of Bayesian and Non-Bayesian Inventory Models,” Management Science. Vol. 30, no. 8, 993-1003

3.Choi, T-M., D. Li, & H. Yan, （2003）“Optimal two-stage ordering policy with Bayesian information updating,” The Journal of the Operational Research Society, Vol. 54, 846-859

4.Chopra, S. & P. Meindl (2002), Supply Chain Management: Strategy, Planning, and Operation, Practice Hall Inc.

5.Eppen, G. & Ananth. V. Iyer, （1997） “Improved fashion buying with Bayesian updates,” Operations Research, Vol. 45 no. 6, 805-820

6.Hunter, N., R. King, & H. Nuttle, (1996) “Evaluation of traditional and quick-response retailing procedures by using a stochastic simulation model,” Journal of the Textile Institute, Vol.87, 42-55