城市级新冠肺炎(COVID-19)疫情预测和仿真模型

City-Level Structured Prediction and Simulation Model of COVID-19

传染病的传播分析对于社会公共安全与城市管理有着重要的意义.针对此次新冠肺炎(COVID-19)疫情的传播,构建了一个城市级结构化疫情预测和仿真模型.该模型基于最新的传染病动力学理论,综合地理信息网络,将易感者、潜伏者、感染者和康复者(SEIR)模型与社会网络模型相结合,对预测区域进行多层级划分;宏观上使用二分网络模拟公共设施与社区节点的关系;微观上使用改进的SEIR模型模拟节点内部的传染情况;节点间使用智能体追踪个体的传播过程.采用国家卫生健康委员会公开的2020年武汉市与北京市疫情确诊及治愈病例数进行对比实验,结果表明,所提模型在疫情仿真上更灵活,预测结果更准确,并能直观地体现不同人群的分布情况和流动趋势.

It is important for social public security and urban management to explore the spread of infectious diseases.A city-level structured prediction and simulation model for COVID-19 is proposed.This model is consisted of SEIR and social network model on the basis of latest infectious disease dynamics theory and real geographic networks.The prediction region is divided into multiple levels.Specifically,a bipartite network is applied to simulate the relationship between public facilities and community nodes at the macro level,and a modified SEIR is applied to simulate the infection within nodes at the micro level.Besides,intelligent agent is applied to track the individual transmission process.The contrast experimental results based on the confirmed and cursed cases of Wuhan and Beijing in 2020 published by National Health Commission,show that the proposed model has better flexibility and higher accuracy,and reflects the distribution and movement of people more directly.