在兩個供貨來源之下探討斷貨風險與成本結構對最適採購批量的影響

Abstract

近十年來的自然災害或人為意外等事件，使得供給中斷開始受到重視，並有了進一步的研究，逐漸討論至斷貨風險的議題。相對於傳統供應鏈管理常探討的需求不確定性，斷貨所造成的供給面不確定性甚是有其研究的價值。本研究將根據過去學者對於供給中斷的研究進行更深入的衍伸，並同時討論斷貨風險以及成本結構，建立本研究決策模型。 本研究乃是以下游零售商的角度，向兩家可能會中斷出貨的供應商進行採購，使用重新報酬理論並以最小化長期總成本做為目標。於決策模型中本研究透過連續時間馬可夫鏈表達供應商經歷正常或維修的程序，並推導出四種情境的移轉機率以及對應之成本。本研究將多項成本皆納入考量，包括採購成本、貨品單位成本、持有成本與缺貨成本等，最後歸結出長期總成本函數。 然而，由於本模型過於複雜，難以確認其特性與最佳解位置，故於此採用了Snyder學者的逼近法，將模型與以化簡，隨後即證明其性質並求得最佳訂購數量與長期平均成本。本研究進一步利用敏感度分析探討各項輸入參數對於成本與訂貨量的影響，並根據分析結果相關管理意涵與建議，以期幫助面臨斷貨風險的廠商決定其訂貨政策。於本研究中發現各項輸入參數的影響性大小與是否為主要採購來源有明顯相關，當主要採購來源的風險增加或成本上升，對於零售商將會產生較高的成本負擔。

In recent decade, the events of natural disasters and human accidents make the issue of supply disruptions become more and more emphasized, which begins to have further research and gradually discuss the topic of disruption risk. Compared to the demand uncertainty studied by traditional supply chain management, supply uncertainty caused by disruptions does have its value to study. This research is based on the model of supply disruptions from the past studies, and the more profound extensions will be conducted in this research. Besides, the effect of disruption risk and cost structure will be discussed simultaneously to construct the decision model in this research. This research is on the standpoint of downstream retailer, who purchasing goods from two unreliable suppliers, using renewal reward theory and set the aim as minimizing the long-run total cost. In the decision model, continuous-time Markov chain is adopted in order to display the process of On and Off period undergone by suppliers, and then figure out the deduction of the transition probability and correspondent cost under four scenarios. Ultimately this research constructs long-run average cost function, which takes ordering cost, unit cost, holding cost and stockout cost into consideration. However, due to the complex of the model, difficult to identify its property and optimum, there this research thus introduces the approximation conceived by Snyder in order to simply the model. Then its property is proved and the optimal order quantity and long-run average cost is figured out. Afterwards, this research uses the sensitivity analysis to understand each the effect of input parameter and proposes relevant management implications and suggestions. The studies discover that the effect of input parameters depends obviously on whether it is main purchasing supplier or not; when the risk or cost increase at the main purchasing source, retailer has to face larger increment of cost.