層級分析法與模糊層級分析法於山坡地生態及災害綜合指標權重分析

A combination of fuzzy and AHP approach to assess the index weight in a GIS-based combined index model for hillside disaster and ecological analysis

郭欣怡(Shin-Yi Guo)

指導教授 ： 朱子偉

國立臺北科技大學/工程學院/土木與防災研究所/碩士(2008年)

https://doi.org/10.6841/NTUT.2008.00205

摘要

層級分析法(AHP)和模糊層級分析法(FAHP)均屬於多評準決策(MCDM)的評估法。層級分析法以階層結構的概念，將複雜且多目標的問題由高層到低層逐步分解，經專家評估後，用特徵向量概念求各指標權重值。層級分析法結構清晰、理論清楚且操作容易，可整合多數決策問題。而模糊層級分析法具有解決現實環境中不確定與模糊性資料之特性，可整合專家對各指標權重之共識。
本研究採用上述兩種方法和模糊理論應用於指標權重分析，重新計算「山坡地生態及災害潛勢綜合指標」模式中各指標相對權重，以綜合評估高雄地區環都市山坡地發生災害與生態受損潛勢。12項指標權重經三種方法分析後，發現皆以坡度、降雨量、土石流潛勢溪流和土地利用四項指標權重，明顯較其他指標高出甚多，顯示綜合專家的評估意見後，此四項指標在災損潛勢的分析中居相當關鍵的地位。在三種分析方法中，應用模糊理論計算指標權重，能更明確兼顧每位受訪者觀點，在本研究中可較完整的呈現各指標的相對重要性。模式以地理資訊系統(GIS)空間分析能力進行整合，計算各指標得分後，以所得綜合指標總分表示災害發生與生態受損之潛勢。模式應用於環高雄地區之山坡地分析，結果顯示高雄市環都市山坡地綜合指標總分，以壽山西南側中山大學後方為最高；此處恰有土石流潛勢溪流通過，土地開發程度亦高，從1986年至今已發生過4起坡地災害；模式因此得以驗證能有效標定較危險區域或易損區域，以達到事先規劃和降低災害的目標。

關鍵字

山坡地 ； 生態 ； 災害潛勢 ； 層級分析法 ； 模糊理論

並列摘要

The Multiple Criteria Decision Making Method (MCDM) means to evaluate asset of alternatives in terms of a number of critera. The Analytic Hierarchy Process (AHP) and Fuzzy Analytic Hierarchy Process (FAHP) are the two major MCDM methods that have recently become increasingly popular. By organizing and assessing alternatives against a hierarchy of multifaceted objectives, AHP provides a proven, effective means to deal with complx decision making. It also assesses the relative importance of multiple choice criteria, which result in a better, easier, and more efficite identification of selected criteria and their weighting as well as analysis. The FAHP, on the other hand, when incorporated with fuzzy logic to AHP, manages to adopt the environmental uncertainty and suggest the relative strength of the factors in the corresponding criteria with fuzzy judgrement matrix to facilitate decision-making. This study aims to determine the relative significance of 12 indexes adopted in a GIS-based combined index model, which has been developed to assess the disaster potential and ecological impairment around hillside areas. Three weighting methods, AHP, FAHP, and modified fuzzy theory integration, are employed in this model, and the result shows that slope, rainfall, debris-flow-prone stream and land use are the four more important indexes than the others. This further concludes expert opinions to consider these indexes to be critical factors in assessing the disaster potential. Compared with the other two weighting methods, the modified fuzzy theory integration method reflects the expert opinions in a more objective way and contributes more to the establishment of the relative importance of 12 indexes. Moreover, the combined-index score produced by model indicates the ecological impairment and disaster potential. When the modified model is applied to hillsides around Kaohsiung city, the highest-scored part in Kaohsiung metropolitan area turns out to be the southwest to Shou Shan and the over-developing backside of National Sun Yat-Sen University, where two debris-flow-prone streams stay nearby and landslides have broken out in the past decades. The model is hereby validated to effectively identify the critical areas with high disaster potential, which will warn the agencies of precaution and remind them of further investigation on those sites by establishing effective engineering measures.

並列關鍵字

Analytic Hierarchy Process ； Fuzzy Theory ； disaster potential ； ecology ； hillside

參考文獻

[36] 翁振益、張淑卿，「資料型態為卜瓦松分配之多屬性決策方法模擬分析比較」，管理思維與實務學術研討會論文集，2004，第1070-1085頁

[33] 邱宏彬，「一個可彈性支援顧客關係管理與資料庫行銷之模糊RFM Model」，電子商務學報，第六卷，第二期，2004，第149-173頁。

[44] 馮正民、邱裕鈞，研究分析方法，建都文化事業股份有限公司，2004。

[22] 王詩怡，台灣地區航空貨運產業發展策略選擇之研究，碩士論文，國立成功大學交通管理學研究所，台南，2003。

[41] 陳育甄，模糊層級分析法應用於城際運具選擇模式之研究，碩士論文，國立成功大學都市計劃研究所，台南，2002。