運用馬可夫決策模式於製造商新製與再製決策

Applying Markov Decision Process in Manufacturing and Remanufacturing Decision

武峙瑄(Chih-Hsuan Wu)

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(2007年)

https://doi.org/10.6342/NTU.2007.00855

摘要

摘要
近年來環保意識抬頭，各國政府立法規範企業應盡的環保責任和應負擔的環境成本；企業內部以節能、減廢、回收再利用等概念重新省視價值鏈上的各項活動；企業和企業間則透過資訊的分享和流程的整並，實踐供應鏈的綠化，以期達到成本降低、利潤分享及環境保護的多重目標。
在廣泛的綠色供應鏈議題當中，再製(Remanufacture)乃是各種逆向物流中使得生產階段所投入的能源和物料被再次利用的一種極有效率的形式。儘管再製造的逆向物流機制目前已廣泛地被企業所採用，但學界在回收再製之議題的研究仍不夠完整；文獻多以線性或非線性確定性模型為分析架構，鮮少針對回收量或需求量為隨機的動態狀況進行分析。本研究運用馬可夫決策模式(Markov Decision Process, MDP)進行模型建構和求解，擬研究一可同時使用全新和回收(Used)物料零件從事生產的製造商，在面對產品需求、回收物料零件的供給皆不確定之情況下，如何做出最小化平均成本的生產決策。本模型僅討論供應鏈中的製造商及回收處理廠，探討製造商採用最佳策略達到穩定狀態(Steady State)後的平均成本和存貨狀況，以歸納最適策略所造成的長期效果。
由研究結果得知，製造商各期的最適再製決策會和該期所處的狀態、回收物料的流失率因子有關。製造商的平均成本則隨再製單位成本、存貨單位成本、回收流失率、最終需求變異的降低而下降，而再製單位成本和回收流失率兩因子的交互效果亦對平均成本產生顯著的影響。因此亟欲降低平均成本之製造商應該在壓低再製單位成本的同時進行回收流失率的降低，以得到成本降低的加成效果。

關鍵字

逆向物流 ； 馬可夫決策模式 ； 新製生產決策 ； 再製生產決策

並列摘要

Abstraction
Environmental issues have got more and more attentions in resent years. Governments started regulating the environmental responsibilities of enterprises. Concepts of energy saving, waste reducing and recovery have been applied in every activity of the value chain inside the enterprises. Green supply chain is also a popular idea which can be achieved through the inter-enterprise information sharing and process reengineering.
Remanufacturing, making most recovery of energy and material, is one of the most efficient form of reverse logistics. Despite that remanufacturing is largely introduced into manufacturer, there is still a lot of space to improve for academic research. An example is that most of the literatures construct linear or nonlinear deterministic models without considering the stochastic pattern of demand. In this research, Markov Decision Process model is used to formulate a cost-minimized production system of a manufacturer, who can produce with both the original material and the used material collected from returned products, facing the uncertainty of demand and returned products. This model focuses on the manufacturer and the recycler; by observing the inventory status under steady states, general conclusion about the impact of critical factors is drawn.
This research suggests that the best production decision under each state is affected by the two state variables and the discarded rate. Remanufacturing unit cost, inventory unit cost, discarded rate, and demand variance have positive impact over average cost of manufacturer, and the effect of interaction between remanufacturing unit cost and the discarded rate has negative influence upon average cost. For a manufacturer trying to reduce cost, lowering the discarded rate is even more important than lowering unit cost.

並列關鍵字

reverse logistics ； Markov Decision Process ； manufacturing decision ； remanufacturing decision

參考文獻

[1] 盧煒竣《納入費率制資源回收機制之製造商成本最佳新製與再製決策模式之研究》，國立台灣大學商學研究所碩士論文，民93。

[11] Fleischmann, M., J.M. Bloembhof-Ruwaard, R. Dekker, Erwin van der Laan, J.A.E.E. van Nunen, and L.N. van Wassenhove, “Quantitative models for reverse logistics: A review,” European Journal of Operation Research, 103, 1, (1997), 1-17

[12] Guide Jr., V. D. R., and R. Srivastava, “An evaluation of order release strategies in a remanufacturing environment,” Computer Operations Research, 24, 1, (1997), 37-47

[13] Guide Jr., V. D. R., “Production planning and control for remanufacturing: industry practice and research needs,” Journal of Operations Management, 18, 4, (2000), 467-483

[14] Hillier, F. S., G. J. Lieberman, “Introduction to Operations Research,” 8th edition, McGraw-Hill, 2005