基于BP神经网络的受灾点的需求紧迫性分级方法  

Classification Method of Demand Urgency of the Affected Points Based on the BP Neural Network

作　　者：姚恩婷 孟燕萍[1] 林国龙[1] 

机构地区：[1]上海海事大学物流研究中心,上海201306

出　　处：《灾害学》2016年第3期211-216,229,共7页Journal of Catastrophology

基　　金：教育部博士点基金项目(20130039;20123121110004);上海市自然科学基金项目(12ZR1412800);上海市科委重点项目(11510501900);上海市科学技术委员会项目(14DZ2280200);上海市教委科研创新项目(13YZ085)

摘　　要：大规模突发事件发生后,往往会涉及到多个受灾区域,使得应急救援的受灾点数目众多。当应急资源有限,运输能力受约束时,为提高应急救援效率,应急物资的调运和配送需要根据受灾点的需求优先级进行。因此,对受灾点进行需求紧迫性的分级排序就非常关键。该文提出了基于BP神经网络的分级方法,构建了影响受灾点的需求紧迫性的评价指标体系,建立了基于BP神经网络的需求紧迫性分级模型。最后,实例验证表明BP神经网络对检验样本的结果输出和期望输出是一致的;并与TOPSIS法、灰色关联分析法和熵权法3种方法的评价结果进行比较,进一步证明了该分级评价方法的科学性和有效性。After the occurrence of large-scale emergencies,often involves multiple disaster areas,so that the number of emergency rescue of the affected point is numerous. When the emergency resources are limited and transport capacity constrained,in order to improve the efficiency of emergency rescue,emergency supplies transportation and distribution needs according to the demand priority of the affected spot. Therefore,hierarchical ordering of the urgent demand of the affected points is critical. This paper proposed a classification method based on BP neural network,constructed evaluation index system of influence the affected points of demand urgency,established the demand urgency classification model based on BP neural network. Finally,example shows that BP neural network for the test sample results output and the expected output is consistent,and it compares the evaluation outcome of TOPSIS method,gray relation method and entropy weight method,and further prove the scientific and rationality of this evaluation method.

关 键 词：BP神经网络 受灾点 应急需求 需求紧迫性 分级 

分 类 号：X43[环境科学与工程—灾害防治]