颱風損害之綜合評估：模糊多準則評估法之應用

Abstract

傳統颱風損害之評估存在兩個主要之課題，第一個課題為傳統對颱風之損害評估多偏向單損害項目與單地區之評估，缺乏跨區域、多項目及能處理不同資料型態的多準則綜合評估方式。另一個課題為颱風損害評估與資料蒐集過程常發生的資訊不充足、不確定與主觀判定之困境。本文之目的在引入模糊理論，建立多準則颱風損害綜合評估模型，以改善傳統颱風損害評估模式的不足。文中以近年台灣六個主要颱風損害為案例，以建立的模式，估計各颱風之綜合損害及各颱風在各縣市的綜合損害，估計結果發現資料所隱含之模糊度愈高，所估計之颱風災害損害程度愈低；相對地，隱含之模糊度愈低則估計之損害程度愈高。代表災害損害估計過程如忽略資料的粗略性，將嚴重扭曲損害估計之成果，可能高估高損害程度之颱風損害，而低估低損害程度颱風之損害。

Conventional evaluation methods on typhoon damage are mainly focused upon mono-item or mono-area, rather than deal with comprehensive damage estimating for multi-items, multi-typhoon or multi-areas. Furthermore, it is usually suffered from the information insufficiency, uncertainty, and subjective judgment while proceeds with the data surveying and damage estimating on typhoon. To the purpose of mitigating these two problems, the main purpose of this paper is to establish a multicriteria comprehensive model using fuzzy set theory, and employ this model as an instrument to evaluate six recently major typhoon damages in Taiwan as a case study. The items in the case study that include two aspects: (1) the comprehensive damage degree evaluating of each typhoon, and (2) the damage geographical distribution estimating in each prefecture/city for each typhoon. It is found that the damage estimation shows where as the potential fuzziness within surveying data increases, as the less of an estimating damages degree, and vice versa. These results represent that if estimator ignores the potential roughness within surveying data in damage estimating process, then the evaluation of damage will be distorted. Therefore, it will be more overestimated the damages, while greater the typhoon damage exposes. In the other hand, will it more underestimate the damage, while which the less the typhoon damages.