基于高分辨率遥感数据的重大自然灾害应急物流系统规划理论研究

Scientific and effective emergency logistics under major natural disasters is an important support to enhance the efficiency of post-disaster emergency response. The existing emergency logistics planning currently has problems such as low precision of emergency demand estimation, weak robustness of emergency facility location design, and poor applicability of emergency transportation routing plan. This project will take the support of national high-resolution remote sensing data in preventing and reducing natural disasters and endeavor to make more meticulous emergency logistics system planning. A mapping model from spatial distribution data to victims’ demand mode will be built to explore an accurate and effective method to estimate emergency material demand. Following the trip chain of emergency logistics activities, i.e., transport emergency materials from pre-positioning facilities to collecting-and-distributing facilities, and transfer to disaster victims, robust emergency pre-positioning and collecting-and-distributing facility location models will be established based on the high-resolution remote sensing data. Some focal issues frequently occurred in the operations of emergency service facilities, e.g., facility disruption, material assignment fluctuation, and the quantitative measurement of victim loss will be captured in the facility location modeling framework. A routing optimization model of describing the emergency transportation between service facilities will be formulated to examine the effect of limited emergency resources and hard time windows on the optimal routing plan. This project will not only expand the application of high-resolution remote sensing data in the area of disaster emergency management but also help understand the generation and evolution mechanism of emergency logistics, and consequently provide support for more scientific emergency logistics system planning efforts.

重大自然灾害环境下科学有效的应急物资供应是提高灾后应急救援效率的重要保障。目前，服务于灾后物资输送的应急物流系统规划仍然存在物资需求预测精度偏低、应急设施选址方案鲁棒性较弱、运输规划实践性不强等问题。本项目拟借助高分辨率遥感数据对防灾减灾战略需求的技术支撑，以应急物流系统精细规划为导向，建立从空间分布数据到应急物资需求样式映射的精准预测模型，获得准确预测救灾物资需求的方法；以应急物资从预置设施至集散设施再转运至灾民的活动出行链为主线，依据高分数据提供的灾情图像信息，建立鲁棒的应急物资预置设施与集散设施选址模型，解决应急物流设施服务失效、资源分配弹性、灾民损失精确度量等问题；通过构建应急物资运输路径规划模型，探索有限的救灾资源与强时间窗要求对运输路径规划方案的影响。本项目不仅能拓展高分数据在灾害应急管理领域的应用，而且有助于深入理解应急物流活动的发生发展过程，为应急物流系统规划提供理论依据。

本项目在前期研究的基础上，基于高分辨率遥感数据提供的灾害背景图像，聚焦考虑灾害环境对应急响应期应急救援与疏散活动产生的影响，从系统整体优化的角度，对应急响应系统中关键节点服务设施选址与运输路径规划等一系列核心理论问题，进行深入精细化的探索性研究，旨在为重大自然灾害应急物资“供应-运输”以及人员疏散与救治规划方案的制定和实施提供更为科学的理论依据。.结合可收集到的高分数据，开发了“城市多系统综合管理应用平台高分数据管理子系统”，实现将卫星源图像包文件进行上传、解压、数据结构化、图像处理、以及分类处理，为项目研究提供所需基础数据；针对高光谱遥感图像中几十乃至上百个连续谱段的光谱数据，在捕获更多细节信息时容易导致维数灾难的问题，研究了基于目标引导的局部约束稀疏子空间聚类的高光谱图像分割算法，以获得更好的分割准确度，为后续数据分析与优化建模的场景设定与条件假设提供依据。.以大规模破坏性自然灾害为背景，结合灾后人道主义应急救援和疏散的活动特征，主要针对应急响应系统中快速提供救援物资以及抢救疏散受灾人员的应急需求，以最小化系统总的运营成本、受灾群体所承受的风险损失以及优化系统整体运行效率等作为优化目标，根据不同的问题场景，开展一系列的应急服务网络优化研究，并设计相应的优化求解算法，以寻求最优的应急响应系统服务策略。.以我国新冠疫情公共卫生事件为研究背景，以大规模核酸检测、医疗物资供应和运输等实际问题为导向，结合新冠疫情下的应急物流及应急活动特征，研究最优的核酸检测服务分配策略、服务路径规划策略，以及疫情期间社区居民应急物资供应的优化分配策略等。.基于上述研究成果，以多个规模的实验路网和济南市部分区域网络为例，开展非常规突发事件背景下应急响应系统服务分配与路径规划方案案例分析，以验证研究成果在实际应用中的适用性。