系统管理学报

系统管理学报 ›› 2022, Vol. 31 ›› Issue (6): 1123-1136.DOI: 10.3969/j.issn.1005-2542.2022.06.009

• • 上一篇    下一篇

中国人口流动网络时空演化及其对新冠疫情传播的影响

石峰1,2，袁韵3，曹炜威4,5，贾建民6,7   

  1. 1. 西南交通大学经济管理学院，成都 610031；2. 服务科学与创新四川省重点实验室，成都 610031；3. 清华大学经济管理学院，北京 100084；4 中国民用航空飞行学院民航飞行技术与飞行安全重点实验室，四川 广汉 618307；5.中国科学院成都计算机应用研究所，成都 610029；6. 香港中文大学(深圳) 经济管理学院，广东 深圳 518172；7. 深圳市人工智能与机器人研究院，广东 深圳 518129
  • 出版日期:2022-11-11 发布日期:2022-11-23
  • 作者简介:石 峰（1992-），男，博士生。研究方向为复杂网络上的人口流动和疾病传播
  • 基金资助:
    国家自然科学基金面上项目(72074072)；国家自然科学基金专项基金资助项目(72042009)；民航飞行技术与飞行安全重点实验室资助项目（FZ2022ZX54）

Spatiotemporal Evolution of Human Mobility Networks in China and Its Impact on COVID-19 Transmission

SHI Feng1,2, YUAN Yun3, CAO Weiwei4, 5, JIA Jianmin6, 7   

  1. 1. School of Economics and Management,Southwest Jiaotong University, Chengdu 610031, China; 2. Service Science and Innovation Key Laboratory of Sichuan Province, Chengdu 610031, China; 3. School of Economics and Management, Tsinghua University, Beijing 100084, China; 4. Key Laboratory of Flight Techniques and Flight Safety, Civil Aviation Flight University of China, Guanghan 618307, Sichuan, China; 5. Chengdu Institute of Computer Application of Chinese Academy of Sciences, Chengdu 610029, China; 6. School of Management and Economics, The Chinese University of Hong Kong (Shenzhen), Shenzhen 518172, Guangdong, China;7. Shenzhen Institute of Artificial Intelligence and Robotics for Society, Shenzhen 518129, Guangdong, China
  • Online:2022-11-11 Published:2022-11-23

PDF

摘要： 基于2020年1月1日至2月29日的手机信令数据构建全国尺度人口流动网络，研究了疫情-春节叠加影响下人群流动行为的时空演化规律，并实证检验了流动网络结构特征在疫情传播中的作用。研究结果表明：大流行对人口流动模式的破坏存在年龄和距离异质性，削弱了高层级城市在人口流动网络中的重要性，并重塑了城市间人口流动强度分布规律。各时间段流动网络度分布均呈现“低度饱和，高度截断”的幂律特征，不同年龄群体迁移模式的城市分异性揭示了不同层级城市居民工作压力存在分化，且分化趋势在年轻人群中更为显著。研究还发现，流动网络拓扑结构与当地疫情传播之间存在因果关联性，采用时间断点回归模型实证检验了流动网络度中心性、接近中心性和特征向量中心性3项结构特征在封控政策对当地疫情传播的影响中起到部分中介作用。本研究对深入理解人口流动网络结构特征在重大公共卫生事件传播中的影响具有重要的理论和实践意义。

关键词: 手机信令数据, 人口流动网络, 新冠肺炎疫情, 时间断点回归, 中介效应

Abstract: Based on large-scale mobile signaling data in China during the COVID-19 outbreak in early 2020, this paper constructs a nationwide mobility network to explore the spatiotemporal evolution of human mobility pattern under the ‘epidemic-Spring Festival’ dual influences and to empirically test the role of the characteristics of the mobility network structure in COVID-19 transmission. The results show that there exist an age and distance heterogeneity in the destruction of the pandemic to human mobility pattern. The pandemic has weakened the importance of core cities in the mobility network and reshaped the distribution law of mobility intensity among cities. The degree distribution in each phase generally follows a power-law distribution, with a low-degree saturation and a high-degree cutoff. The city heterogeneity of migration patterns among different age groups reveals that the work stress of residents in different cities is differentiated, which is more prominent in the young population. The results also demonstrate that there is a causal relationship between the topology of the mobility network and local epidemic transmission. The regression discontinuity in time is used to examine the partial mediation effects of degree centrality, closeness centrality, and eigenvector centrality in the impact of quarantine policies on the local COVID-19 transmission. This paper provides a better understanding of mobility networks structure in determining the spread of major public health incident, which has important implications in theory and practice.

Key words: mobile signaling data, human mobility networks, Covid-19, regression discontinuity in time, mediation effects

中图分类号: