多目標決策分析結合約略集合理論探討供應鏈網路決策問題

Multi-objective decision making analysis combined with rough set theory in supply chain network decision making problems

羅懷暐(Haui-Wei Lo)

指導教授 ： 劉建浩

國立臺北科技大學/管理學院/工業工程與管理系碩士班/碩士(2013年)

https://doi.org/10.6841/NTUT.2013.00035

摘要

決策分析在供應鏈網路問題上通常不只有單一目標，例如選擇供應商時經常希望以最小化總成本與最大化服務水準，為目標進行規劃。如何滿足這些條件並達到各目標的渴望水準，以及掌控供應鏈問題的投入與產出等各種決策變數，是多目標決策問題。過去的這類問題透過各種演算法可得多個可行解，以提供決策者選擇，而決策者再依各自偏好或經驗挑選出理想方案。本研究則以資料探勘的方式，在諸多方案中建構決策規則，使每次決策品質有所依循而穩定，也可達到知識管理的目的。本研究探討供應鏈網路問題目標函數為整體成本最小化與供應鏈可靠度最大化，同時考慮產能、供應與生產平衡、顧客需求等限制，先以基因演算法求得多個可行解，再以集群分析將這些解分組，最後由約略集合理論方法(Rough Set Theory)產生決策規則，形成知識庫，引導決策者做出最符合實際的決定。此方法的結合使得供應鏈網路系統隨著需求變動而改變，且供需更加穩定，決策規則方便決策者有參考依據，讓評選方案更加客觀。為證明本研究所提出之模型，本論文模擬某精密機械裝配廠，實際調查相關資料，建構網路供應鏈模型，所獲得之結果證明此方式為一實用且有效的方法。

關鍵字

供應鏈 ； 多目標規劃 ； 集群分析 ； 約略集合理論 ； 基因演算法

並列摘要

In general, there are more than one goal in supply chain management (SCM); such as,
total cost minimization and service level maximization for target planning when
choosing a supplier. There may be many constraints among suppliers, manufacturers,
distributors and customers in supply chain network analysis; however, how to meet
these conditions, to achieve each goal with an ideal level and to control input and
output of supply chain network as well as other decision variables are the
multi-objective decision problem. In the past, several feasible solutions of this kind of
problems could be provided for decision maker to choose from a set of solutions. A
decision maker could choose an optimal solution based on personal preference and
experience. Nevertheless, this kind of decision may vary from person to person. In
order to make each decision quality has rules to follow and stable, this study applies
data mining and derives some decision rules among a set of solutions. Moreover, the
iii
method could also become the base of knowledge management. The objective functions
of the supply chain network design are to minimize of total cost and maximize the service levels. Capacity, balance of supply and demand, customers’ requirement and other constraints are also taken into consideration. To help a decision maker to make the most practical decision, this study applies the Genetic Algorithm to obtain a set of feasible solutions. Then, we use the Cluster Analysis to categorize these solutions and generate the decision rules through Rough Set Theory. This integrated method considers various variables and makes supply and demand more stable. Based on decision rules, a decision maker could make a decision easily and objectively. In the study, a precision machinery assembly company is chosen as an example. According to this real data and supply chain network design, the empirical case study shows the model is practical and effective.

並列關鍵字

Supply Chain ； Multi-Objective Decision ； Cluster Analysis ； Rough Set Theory ； Genetic Algorithm

參考文獻

郭瑞祥、蔣明晃、陳亞男、劉基全 (2004)，「多廠生產規劃之供應鏈決策支援模式」，臺大管理論叢DEC，第15卷，第1期，第49-74頁。

藍俊雄、邱誌偉、胡鴻慶 (2007b)，「整合式多期規劃供應鏈模式」。科學與工程技術期刊，第3卷，第4期，第87-98頁。

Ahmadi, M., Yousefi, A., Soroudi, A., Ehsan, M. (2008), “Multi objective distributed generation planning using NSGA-II”, Power Electronics and Motion Control Conference, 13: 1847-1851.

Al-e-hashem, S. M. J. M., Malekly, H., Aryanezhad, M. B. (2011), “A multi-objective robust optimization model for multi-product multi-site aggregate production planning in a supply chain under uncertainty”, International Journal of Production Economics, 134 (1): 28-42.

Amin, S. H. and Zhang, G. (2012), “A multi-objective facility location model for closed-loop supply chain network under uncertain demand and return”, Applied Mathematical Modelling, In Press, Corrected Proof-Note to users.