模糊方法在質化評估準則之運用--以交通計畫土地徵收難易度為例

Abstract

傳統質化準則常以「主觀等級給分」的方式來衡量，其過於主觀且忽略本 身亦可能具有多項指標可用以衡量該質化準則。基於此，如何讓質化準則 由「主觀」的陳述轉變為「客觀」的衡量？如何彙總不同評判者的認知？ 如何整合質化準則本身內含的多個不同的指標？...等等問題，即為本研 究欲探討的課題。本研究運用模糊方法針對單一質化準則之評估值的衡量 做深入的探討，評估模式的建立步驟如下：首先透過「語意變數法」求得 各「初選」指標之權重，依此進行「決選」指標的選取；其次將各決選指 標對質化準則的影響劃分為數個程度等級，每一個等級代表一個模糊子集 ，透過「模糊統計試驗法」建立各決選指標影響質化準則所屬各模糊子集 的隸屬函數；最後根據模糊識別之「最大隸屬度原則」與「覆蓋性」觀念 ，進行各項指標的整合工作。本研究以「土地徵收難易度」質化準則為例 進行實例研究，問卷對象為曾從事交通建設土地徵收工作之政府公務員， 並依上述步驟建立評估模式。首先選取七項初選指標，分別為私有地％、 都市計畫區％、土地價格差距％、建物面積％、總地主人數、地主年齡與 地主學歷，經指標權重大小評選後，以前五者為決選指標。本例中將各決 選指標對質化準則的影響都劃分為「容易」、「尚可」與「困難」三個程 度等級，分別建立各決選指標所屬各子集的隸屬函數。最後將各項指標加 以整合，即可建立較具客觀性的「土地徵收難易度」之衡量值。在模式驗 證分析中，以本整合方法與「模糊綜合評判法」進行模式應用績效的比較 ，根據問卷調查結果，發現本整合方法有較高的正確判斷率，證實本研究 在此實例中有較高的可信度。 Tranditional qualitative criteria evaluation is usually subjective and ignores there could have many indicators to measure it. This study aims at exploring important issues such as how to change qualitative criteria evaluation from subjective statement to objective evaluation, how to aggregate different decision maker's evaluation results, and how to combine many different indicators that measure the qualitative criterion.In this paper,a qualitative criteria evaluation algorithm based on the concepts of fuzzy set theory is proposed. There are several steps to establish the evaluation model. First, possible indicators are considered in evaluating the qualitative criterion. The linguistic variable method is then used to calculate the weights of indicators, and part of the indicators are chosen as the final evaluation indicators by the weights. Second, the chosen indicators are categorized into several fuzzy groups, the membership function of each fuzzy group is established by using fuzzy statistics method. Finally, according to fuzzy indentification and coverage principle, we combine all evaluation indicators to a composite result.The measurement of difficulty level of land acquisition is taken as a case study. To do this, we survey 40 government officials who undertake the land acquisition work in transportation projects. First, we select seven indicators which are private land percentage, urban planning percentage, land price gap percentage, building area percentage, total number of land owners, land owner's age and the education level of land owner. The first five are chosen by examining their weights to evaluate the qualitative criterion. Three fuzzy groups, "easy" , "fair" and "difficult",are categorized for each chosen indicator. Second, we construct a membership function to each fuzzy group of each indicator. Finally, we combine all the indicators to a more objective measurement. To compare the results between fuzzy synthetic decision method and the method in this study, we design a survey to ask the government officials and the results shows the the method in this study could be better in term of accuracy.