消耗性救災物資需求預測

Abstract

救災物流輸配送系統在災害發生後不但需肩負救援物資之輸送，更應適時、適地、適量的將救援物資送抵災區，以支援救災行動之進行。且因救援時間、救援物資的質與量、災區資訊均有限，當大規模地震災害發生後，救災單位若能化被動為主動，在短時間內統整災區資訊、預測災區物資需求，以供決策者迅速、有效地分配救援物資，則可有效避免物資配送不及或供需失衡等現象。本研究以消耗性救災物資需求為例，建構消耗性物資需求預測模式。 鑒於大規模災害發生初期，災區往往發生資訊傳遞不易或災情消息紊亂之情形，因此第一階段模式首先透過資料融合之程序，降低各可得消息來源公佈死亡人數資料不確定性，使決策者能確切掌握災區情形；其後透過物資需求量推估模式，以資料融合後之死亡人數資料推估災民所需物資數量。第二階段則針對救災物資供需不均之情形，透過定義質化與量化需求屬性，衡量各災區對於物資需求之急迫程度，且以類神經網路模式將各災區之物資需求依其屬性聚類分群，並使用TOPSIS法將各群組依其對物資需求急迫程度排序，以作為物資配送順序之參考。由於實証資料取得不易，因此使用模擬資料進行數值測試與情境模擬分析，以評估模式適用性。模擬結果顯示，本研究之消耗性救災物資需求預測模式，於實務應用具有可行性。

Because of rescue time, quantity and quality of resources and information of affected areas are limited. After the large-scale earthquake disaster happened, how to distribute urgent relief effectively, efficiently and precisely is vital to the alleviation of disaster impact in the affected areas, which remains challenging in the field of logistics and related study areas. In this study, we present a data-fusion approach to the operations of emergency relief distribution system responding to deal with the disorder information in the initial stage of disaster. Based on a proposed 2-layer consuming emergency relief demand forecasting conceptual framework, the proposed methodology involves two recursive mechanisms: (1) consuming relief demand forecasting, and (2) disaster-affected area grouping. Numerical studies with a simulated data sets are conducted, and the corresponding results indicate the applicability of the proposed method and its potential advantages. We hope that this study can not only make the proposed emergency logistics system available with more benefits to the development of emergency logistics systems for the urgent needs of disaster areas around the world but also stimulate more excellent researches concerning emergency logistics management.