陸用飛彈裝備零件供應商之評選—應用DEA與PCA方法

An Integrated Supplier Selection Model for the Components of Missile Equipments Used in the Army

林欣毅(Hsin-Yi Lin)

指導教授 ： 徐旭昇教授

元智大學/工程學院/工業工程與管理學系/碩士(2008年)

摘要

國軍飛彈裝備大部分均仰賴國外廠商製造、生產，然因修護用料採購不及與運補時程冗長，導致國內裝備維修效益不彰，影響部隊戰備任務遂行。近年，因國內各型企業電子、電機、機械五金之製造、維修水平已與國外廠商相近，另一方面為了扶持國內產業發展，在經濟部及國防部共同指導下，將維修所需用料釋放國內廠商試製（修），希望集結民間力量及產業共同發展，秉持「凡民間有能力供應者，國軍不建能量，也不對外採購」之原則，結合民間企業，在國內建立製造（維修）的商源，縮短送美回修時程，大幅提昇裝備妥善，以創造雙贏契機。
自從實施軍品釋商政策以來，每年參與試製（修）廠商多達數百家。本篇論文主要目的為發展一種整合資料包絡法（DEA）、主成份分析法（PCA）與多屬性效用函數權重評估法的模式來對承製（修）廠商之績效實施評估。我們審慎選擇其中20家廠商做為研究樣本，先將主管最關心的產品價格、壽限、保固等三項因素排除在外，將其餘重要因素建立評選架構。架構分為2層，第1層評估項目為投入與產出，其中投入項包含資源及成本等2項，而產出則為效益、品質等2項；第2層則將資源再分為員工數及資產總額等2項，成本則分為研究發展費、設施投資及原物料成本等3項，效益則分為營業收入淨額、稅後淨利、營收成長率等3項，品質則分為交期準確度及產品可靠度等2項。過程分兩步驟：計算(1) DEA之AP效率及計算(2)結合DEA與PCA模式效率－下層項目評估方式為DEA之AP效率，而上層則使用PCA。從兩模式評比排序找出最優之數家廠商，再依主管最關心三項因素建議使用多屬性效用函數權重評估法進行最後之裁決。
經過計算後排序，使用Spearman Rank及Kendall’s Tau兩種無母數統計方法對此兩種模式之排序進行相關性驗證，在此個案中證明此二種排序為高度正相關。

關鍵字

資料包絡分析法 ； 主成份分析法 ； 多屬性效用函數 ； 無母數統計分析

並列摘要

This thesis studies a real life case for supplier selection arising in a military organization that governs missile equipments, and proposes an integrated multi-criteria decision making model, which combines the techniques in data envelopment analysis (DEA), principal component analysis (PCA), and multi-attribute utility function decision theory, to help the management in the military organization select the most proper supplier for their equipment components. The decision making method consists of three phases: (1) Analyze and classify the important attributes in evaluating suppliers; (2) Understand the attributes that the management are most concerned with; (3) Construct a two-level hierarchical model for all attributes not in (2); (4) Select several most competent suppliers based on the model built in (3), using the AP efficiency rate in DEA and the efficiency rate in the integrated model combining methods in DEA and PCA; (5) Determine the best supplier using a multi-attribute utility function decision making method.
In the case study, 20 qualified suppliers with relevant data were first selected. The developed integrated model is then applied to solve the supplier selection problem. At the fourth step, four competent suppliers were selected and the final decision was made based on the weighted sum of the utility values of the three attributes that the management is most concerned with. In the case, the three attributes are prices, life times, and warranty length.
Additionally, Spearman rank test and Kendall’s Tau test are employed to compute the correlation between the supplier rankings computed by the AP model and the DEA-PCA model. Both test results indicate that the ranking produced by these two models are highly positively correlated.

並列關鍵字

Data envelopment analysis ； Principal component analysis ； Multi-attribute utility functions ； Non-parametric statistical testing

參考文獻

15.徐基生等，「運用資料包絡分析法評量工業技術研究院各研發組織之研發績效」，管理評論，第二十二卷第二期，頁25~53，民國九十年。

2.廖麗卿，「應用AHP與DEA模式評選供應商之研究—以戰、甲砲車夜視裝備為例」，元智大學工管研究所，碩士論文，民國九十四年。

18.洪海玲，「以資料包絡分析法作製造業之營運效率分析」，國立成功大學工業管理研究所，碩士論文，民國九十一年。

19.徐仕華，「資料包絡分析法之應用－台灣光電產業跨週期效率衡量」，靜宜大學企業管理研究所，碩士論文，民國九十三年。

1.Adler N, Berechman J. “Measuring airport quality form the airlines view point: an application of data envelopment analysis,” Transport policy, 8, pp. 171-181, 2001.