以群體決策理論解析採購評選作業之研究

Analyzing Procurement Evaluation by Group Decision Making Theory

系所名稱：工程科技研究所 學位別：博士 畢業學年：96年 指導教授：王隆昌 我國於1998年制訂政府採購法時，納入國際間通用的最有利標決標制度，期望能將採購競爭重心由純粹的價格競爭，移轉至品質及服務。實施迄今，雖然在若干案例頗受肯定，但亦有不少爭議與質疑。本研究透過檢討法令規定、回顧學界理論、蒐集實際案例及分析調查結果等方式，探討目前採購評選過程中值得注意的之爭議焦點。研究發現，現行評選委員之任務與權限，缺乏一致明確的規定。另外，現行法令雖然規定，當評選結果產生明顯差異時，召集人應進行處理。但該規定因缺乏客觀認定「顯著差異」之標準，實務上難以執行。此外，不但個別委員的評選結果彼此需進行比較，且個別委員的評選結果亦需與工作小組的初審意見進行比較，程序過於繁複，反而不易界定產生差異的根源。最後，當評比結果產生明顯差異後，現行法令規定有四種處理方式。但該四種處理方式對於原評選結果均造成不同程度的衝擊，目前卻未規範適用時機。本研究利用新制度經濟學、方案評選等學理，建議應於政府採購法中明確規定，得標廠商最終核定權歸屬於機關首長，藉由首長對於評選程序之認可核定權限，設定評選委員會評定結果之最後檢驗機制。但相對地，若機關首長欲退回評選委員會之評選結果，應書面敘明意見及理由，以避免機關首長恣意專斷。有關招標文件中之評選項目、評審標準及評定方式，則應回歸由各機關自行依採購目標制訂。評選過程之重要程序事項，亦應回歸由機關負責控制。此外，本研究結合相對優勢比較矩陣及集群分析，提出適於現行法令規定採用的評選差異識別程序。該識別程序，不僅合乎人性判斷、具備學理基礎，且能檢驗全體評比結果是否一致，並指出產生差異的根源。由實際案例分析結果可知，此識別程序確實可以明確量化的方式，辨識委員間評選結果的差異情形，對於最有利標之評選品質具有實質提升的功能。

At the time when Taiwan government constituted the Government Procurement Act (GPA, hereafter), the “Most Advantageous Tender” (MAT, hereafter), the most commonly used system around the world was led in, which was expected that the procurement competition core would transfer from pure price competition to quality and service level. So far, though there were many successful cases, many disputes and challenges existed as well. This research intends to discuss the controversial issues worth to be noticed during the procurement evaluation.through different ways, such as examining regulation statements, reviewing academic theories, gathering practical cases, and analyzing survey results. The findings show that the current task and jurisdiction of committee members are lacking consistent and clear-out rules. In addition, though it is stipulated that when there exists significant discrepancy in evaluation results, the chairperson should deal with it immediately. However, this stipulation is hard to execute in practice for a lack of objective standard to determine the “significant discrepancy.” Besides, not only the results of individual committee members needs to be compared with each other, but their results also have to be compared with the original opinion of working group. The overly complicated processes make it a tough task to define the sources of discrepancy. Finally, once there exists significant discrepancy in evaluation results, there are four approaches defined in the current legislation to deal with it. All the four approaches have different degrees of impact on the original evaluation results; however, the appropriate timings for using them are not defined yet in the related regulations. This research utilizes theories of the new institutional economics and project evaluation and recommends that the GPA should define clearly that the final right of determination belongs to the director of the entity. Through the right of determination, the director of the entity is set to be the last verifying mechanism for the results of the committee. However, if the director would like to change the decision of the committee, the director should deliver his opinions and reasons in order to avoid arbitrariness. As for the evaluation items, judging standards and evaluation approaches listed in the tender documentation, they should all be determined by each entity according to their procurement goals. The important procedural regulations should also be under control of the entities. Additioally, this research combines the comparative-advantageous matrix and cluster analysis to set forth the suitable discrepancy identification process for current legislation. Such process not only meets with the human intuition and theoretical assumptions, and it could also be used in verifying the consistency of overall result of evaluation as well as pointing out the sources of discrepancy. From practical case studies, this proposed identification process is proved to practically improve the quality of the MAT in that it uses quantitative approaches to identify the discrepancy among evaluation results made by the committee members.