基于范例推理的灾害性地震应急物资需求预测研究

Research on Predicting Emergency Material Demand after Disastrous Earthquake Based on Case-based Reasoning

基于范例推理(CBR)理论,利用最近相邻法和粗糙集理论搜索相似度最高的主震历史范例,分析各主要物资需求量的影响因素,预测当前范例主震期应急物资需求量。通过序贯决策,采用马尔科夫预测模型预测余震类型,进而搜索余震历史范例,预测余震期应急物资需求量。以"玉树"地震为例,运用该方法估算地震发生后食物类、生活用品类、药品类、工程机械类的需求量。

For a presumed present disastrous earthquake, the emergency material demand was predicted as mentioned below：highest similarity historical examples of it were searched out using Nearest Neighbor Algorithm and Rough Set knowledge based on the CBR theory; affecting factors on the demand were analyzed ; the main demand of emergency material after a principal seismic stage was predicted ; the aftershock type was predicted by sequential decision-making and adopting Markov forecast model, aftershock examples were searched out and the emergency material demand of aftershock was predicted. Taking Yushu earthquake as an example, the method was used to predict the demands of foods, living goods, drags and mechanical engineering things after earthquake.