DEA模型應用於台電配電技術部門績效評估之研究

Abstract

DEA模型應用於台電配電技術部門績效評估之研究 研究生：龔良智 指導教授：張保隆 國立交通大學經營管理研究所 摘 要 本研究以台電公司現有之統計資料為基礎，利用資料包絡分析法（Data Envelopment Analysis）衡量台電公司台灣省21個區營業處配電技術部門經營績效（performance）。其主要目的如下：一、發展一個包含多元指標的績效評估衡量模式。二、以DEA方法應用於發展之效率衡量模式上，求出21個區營業處配電技術部門之整體效率（aggregate efficiency）、技術效率（technical efficiency）及規模效率（scale efficiency）並加以解釋分析。三、以差額變數分析（slack variables analysis）探討各個配電技術部門對於資源使用狀況及可能改善的方向與幅度。四、以敏感度分析(sensitivity analysis)探討投入項及產出項項目數改變時，對效率衡量結果的影響程度。 實證結果顯示在21個區營業處配電技術部門經營績效上，整體相對有效率的僅有二個，相對效率介於1及0.9間的有八個。整體績效不彰的單位，部分係由技術效率偏低，可藉由投入資源之配置與運用分析檢討改善；部分則為規模效率不佳所致，可經由規模報酬分析求出其應改善之方向。且由差額變數分析顯示，工務類大多為投入過高，尤其在機具設備成本方面，產出要素影響較小；而設計、維護、調度類則多為產能不足，投入要素的影響反較小。敏感度分析在工務類方面發現去除某一投入或產出項時原相對效率大多較初始值低，表示其邊際貢獻呈遞增，宜增加該要素使用量。另研究結果也顯示相對效率值的高低，與地區位置、城市或鄉村等特性無關。甚至藉由相關係數分析，維護與調度類相對效率值的高低亦與該部門學歷無甚高之相依關係。 從管理階層來看，面對技術效率不彰之單位可透過參考集合提供單位主管在提昇技術效率時一個模仿之對象；對於規模效率不佳單位可藉由本研究所繪之規模報酬趨勢圖看出各單位內部及各單位間調整之優先趨勢，準此，均可作為決策者或主管機關未來調整時之參考。

Performance Evaluation of Taipower's Distribution Technical Sectors with Data Envelopment Analysis Student：Liang-Chih Kung Advisor：Pao-Long Chang Institute of Management Science National Chiao-Tung University Abstract Based on Taipower's statistical data and Data Envelopment Analysis （DEA）,this study aims to evaluate the performance of distribution technical sectors in Taipower's 21 district offices. There are four purposes in this study: 1) to develop a multi-index performance evaluation model , 2) to develop an efficient evaluating model and to obtain aggregate , technical and scale efficiency of the distribution technical sectors by using DEA model as well as to explain the reasons, 3) to explore the resource application in every distribution technical sector and to find out the improving countermeasures, 4) to explore the variation of result efficiency with sensitivity analysis, when the amount of input or output factors change. The empirical result of managerial performance of distribution technical sectors shows that the aggregate efficiency is 1 in 2 offices, and between 1 and 0.9 in 8 offices. Some offices have poor overall performance due to: 1) low technical efficiency that can be improved by making best of resource allocation analysis, 2) poor scale efficiency , which can be improved through returns-to-scale analysis. The result of slack variable analysis shows that 1) the input is too high ,especially the cost of machine equipment in the Construction sector, while the influence of output is relatively small, 2) the output is inadequate in the Design, Maintenance and Dispatch sector, while the influence of input factor is small. The sensitive analysis shows that the aggregate efficiency is smaller than the initial values if the input or output factor in the Construction sector decreased. This means that there is an increasing marginal contribution, therefore, the amount of the factors should be increased. The results of the study show that there is no connection between the location of the offices and the aggregate efficiency. Furthermore, correlation analysis also shows that there is little relationship between the relative coefficient value and employee's educational background in the Maintenance and Dispatch Sectors. From the managerial point of view, the sectors that have low technical efficiency can improve their efficiency by learning from other offices in the reference set. While those that have low scale efficiency can obtain the priority of intra- and inter-sector coordination through returns-to-scale trend figure represented by this study. The result of the thesis can be helpful to decision makers or managers for future organizational re-engineering.