支持灾难救援的在线空间众包匹配算法  

Online Spatial Crowdsourcing Matching Algorithm for Disaster Relief

作　　者：刘俊岭[1,2] 吴晴晴 董珊珊 孙焕良 许景科[1,2,3] Liu Junling;Wu Qingqing;Dong Shanshan;Sun Huanliang;Xu Jingke(School of Computer Science and Engineering,Shenyang Jianzhu University,Shenyang 110168,China;Liaoning Province Big Data Management and Analysis Laboratory of Urban Construction,Shenyang 110168,China;Shenyang Branch of National Special Computer Engineering Technology Research Center,Shenyang 110168,China)

机构地区：[1]沈阳建筑大学计算机科学与工程学院,辽宁沈阳110168 [2]辽宁省城市建设大数据管理与分析重点实验室,辽宁沈阳110168 [3]国家特种计算机工程技术研究中心沈阳分中心,辽宁沈阳110168

出　　处：《南京师大学报（自然科学版）》2024年第4期21-30,共10页Journal of Nanjing Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(62073227);国家重点研发计划课题(2021YFF0306303);辽宁省教育厅资助项目(JYTMS20231596、LJZ2021008)。

摘　　要：灾难发生后人们常常通过社交媒体平台发布志愿者救援与受灾者求助信息,从这些数据中抽取求助任务与救援人员信息并对两者进行合理匹配可以为救助提供有效支持.本文将空间众包技术引入灾难救援领域,提出支持灾难救援的在线空间众包匹配问题.利用深度学习分类方法与大规模语言模型构建灾难事件信息抽取模型,实现了救援和求助信息的准确抽取;设计了任务等级评定方法与动态损失度量,以反映任务的紧急性和损失的动态变化;基于动态损失度量提出了一种综合抢占与延迟策略的贪心算法.通过真实数据集及合成数据集进行详细的实验分析,与现有算法相比,提出的综合抢占与延迟的贪心算法总损失至少减少35%,验证了所提算法的有效性.After disasters,people often post information about volunteer rescue efforts and requests for help from the affected on social media platforms.Extracting the information of help task and rescue personnel from these data and making a reasonable match between them can provide effective support for rescue.In this paper,spatial crowdsourcing technology is introduced into the field of disaster relief,and online spatial crowdsourcing matching problem for disaster relief is proposed.The disaster event information extraction model is constructed by using deep learning classification method and large-scale language model to realize the accurate extraction of rescue and help information.The task rating method and dynamic loss measurement are designed to reflect the urgency of the task and the dynamic change of the loss.A greedy algorithm combining preempt and delay strategies is proposed based on dynamic loss measurement.Through detailed experimental analysis of real data sets and synthetic data sets,the total loss of the greedy algorithm combining preempt and delay strategies is reduced by at least 35% compared with the existing algorithm,and the effectiveness of the proposed algorithm is verified.

关 键 词：空间众包 灾难救援 任务匹配 任务等级 最小损失 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]