基于ANP-RF算法的泥石流严重度评估  

Debris Flow Severity Assessment Based on ANP-RF Algorithm

作　　者：王艳锦 武晓春[1] 郭荣昌[1] WANG Yan-jin;WU Xiao-chun;GUO Rong-chang(School of Automation and Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学自动化与电气工程学院,兰州730070

出　　处：《兰州交通大学学报》2021年第1期66-72,共7页Journal of Lanzhou Jiaotong University

基　　金：国家自然科学基金(61661027)。

摘　　要：针对泥石流造成的经济损失和人员伤亡问题,提出一种基于ANP-RF算法的泥石流严重度评估方法.首先建立多准则、多层次的ANP结构模型,计算风险因素的主观权重,并给予RF算法历史数据进行训练、测试,得出风险因素的客观权重;再根据主、客观权重计算组合权重;最后加权计算组合权重和各因素数据量,得到泥石流灾害的严重度等级.以川藏线成康铁路线路为例,选取线路类型、坡度、土地利用类型、土壤类型、年平均降雨等11个相关因素进行成康铁路沿线泥石流灾害严重度评估.评估结果可为泥石流灾害的防治、信号集中监测系统预警等管理工作提供参考.In view of the economic loss and casualties caused by debris flow,this paper puts forward a method to evaluate the severity of debris flow based on ANP-RF algorithm.Firstly,the multi criteria and multi-level ANP structure model is established to calculate the subjective weight of risk factors,the RF algorithm is used to train and test the historical data to get the objective weight of risk factors.Then,the combination weight is calculated according to the subjective and objective weight.Finally,the severity grade of debris flow is obtained by the weighted sum of the combined weight and the data quantity of each factor.Taking Chengkang railway of Sichuan-Tibet Railway as an example,11 related factors such as line type,slope,land use type,soil type and annual average rainfall are used to evaluate the severity of debris flow.The assessment result can provides reference for the prevention and control of debris flow disaster,the early warning of centralized signal monitoring system and other management work.

关 键 词：ANP-RF 严重度评估 组合权重 泥石流 

分 类 号：U216.419.2[交通运输工程—道路与铁道工程]