滑坡灾害应急避难所两阶段选址布局  

Two-stage Site Layout of Landslide Disaster Emergency Shelter

作　　者：赵佳佳 郭海湘[1,2,3] 黎金玲 顾明赟 左芝鲤 王子夏 ZHAO Jia-jia;GUO Hai-xiang;LI Jin-ling;GU Ming-yun;ZUO Zhili;WANG Zi-xia(School of Economics and Management,China University of Geosciences,Wuhan 430074,China;China Mineral Resources Strategy and Policy Research Center,China University of Geosciences,Wuhan 430074,China;School of Management,Xi'anUniversity of Financeand Economics,Xi'an 710100,China;School of Management,Hubei Universityof Education,Wuhan 430205,China;School of Economics and Management,Xinjiang University,Urumchi 830000,China;School of Economics and Management,Wuhan University of Engineering Science,Wuhan 430200,China)

机构地区：[1]中国地质大学经济管理学院,湖北武汉430074 [2]中国地质大学中国矿产资源战略与政策研究中心,湖北武汉430074 [3]西安财经大学管理学院,陕西西安710100 [4]湖北第二师范学院管理学院,湖北武汉430205 [5]新疆大学经济与管理学院,新疆乌鲁木齐830000 [6]湖北工程科技学院经济与管理学院,湖北武汉430200

出　　处：《系统工程》2022年第5期140-149,共10页Systems Engineering

基　　金：国家自然科学基金资助项目(71874165,72074198);国家社科基金重点项目(21AZD074);中央宣传部“宣传思想文化青年英才”项目(2020084007);全国统计科学研究项目(2020LY007);湖北省重点研发计划项目(2021BCA219);中央高校基本科研业务费专项资金项目(CUG2642022006);湖北省教育厅科学研究计划项目(B2019311)。

摘　　要：为提高滑坡灾害多发区疏散效率,针对滑坡灾害特点提出了滑坡点评价-避难所选址的两阶段方法。首先,根据研究区实际情况,选取滑坡的13个特征指标,利用BP神经网络对滑坡隐患点进行危险性评价,识别危险性较大的滑坡点,并将其作为避难所选址布局的重要依据之一;其次,考虑避难所数量最小、避难所与滑坡点距离最小值最大、避难所与受灾点距离最大值最小等目标,建立避难所多目标选址模型,利用NSGA-II算法求解模型的Pareto最优解集,最后选取解集中总疏散距离最小的方案作为最优方案。本文以三峡库区湖北省秭归县郭家坝镇为例验证方法的有效性,结果表明该研究方法能够将滑坡灾害的发生与避难所的选址结合起来,有针对性地疏散滑坡高危险区内居民,有效提高了疏散效率。Landslide disasters are the most frequent geological disasters in mountainous areas in China. Scientifically and rationally setting up emergency shelters can effectively evacuate the affected people quickly. In order to improve the evacuation efficiency in landslide hazard-prone areas, this paper proposes a two-stage method for landslide point evaluation and shelter location selection according to the characteristics of landslide hazards. First, select 13 characteristic indexes such as slope, aspect, plane curvature, profile curvature, surface runoff coefficient, sediment transport index, soil moisture, vegetation, land use, road density, river density, lithology, fault density, etc. The BP neural network evaluates the landslide hazard to hidden danger points of landslides, identifies the more dangerous landslide points, and uses them as one of the important basis for the layout of the shelter site. Second, consider the minimum number of shelters, the maximum distance between the shelter and the landslide point, and the distance between the shelter and the disaster point. The target with the maximum value is the smallest, the multi-target location model of the shelter is established, the NSGA-II algorithm is used to solve the Pareto optimal solution set of the model, and finally the solution with the smallest total evacuation distance in the solution set is selected as the optimal solution. This paper takes Guojiaba Town, Zigui County, Hubei Province, Three Gorges Reservoir for example to verify the effectiveness of the method. The results show that the research method can combine the occurrence of landslide disasters with the location of shelters to evacuate high-risk areas of landslides in a targeted manner. The residents inside have effectively improved the evacuation efficiency.

关 键 词：滑坡 两阶段选址 避难所 多目标模型 BP神经网络 NSGA-II算法 

分 类 号：X43[环境科学与工程—灾害防治]