PROGRESS IN GEOGRAPHY ›› 2021, Vol. 40 ›› Issue (12): 2025-2034.doi: 10.18306/dlkxjz.2021.12.004

• Articles • Previous Articles     Next Articles

Formation mechanism of network compactness in Jiangsu Province based on exponential random graph models

HAN Gang1(), SHI Xiusong2,*(), LIU Zhimin3   

  1. 1. Faculty of Architecture and Civil Engineering, Huaiyin Institute of Technology, Huai'an 223001, Jiangsu, China
    2. North Jiangsu Development Academy, Huai'an 223001, Jiangsu, China
    3. School of Geographic Sciences, East China Normal University, Shanghai 200241, China
  • Received:2021-01-13 Revised:2021-07-01 Online:2021-12-28 Published:2021-12-24
  • Contact: SHI Xiusong;
  • Supported by:
    National Social Science Foundation of China(19BJY001);Social Science Foundation of Jiangsu Province(18EYC003);Social Science Research Project of Huaian(2021C39)


Under the background of information technology progress, urban and regional research has begun the shift to the relationship between network organization function and structure, which is formed by the factor flows. Numerous studies indicate that it is feasible for urban network to simulate the urban region system relationships with the abstract pattern of regional structure. The networked urban region system serves as the precondition to explore the compactness of regions and provides a feasible quantitative research approach for the analysis of the compactness characteristic of regions and the mechanism. This research established the concept of network compactness, which means the manifestation of region compactness in the spatial structure of the urban network. In the context of urban network organization, it refers to the collaborative and complementary relationship among cities of all levels in the network space. With the purpose to explore the influencing factors of this relationship, empirical research was conducted by taking districts and county-level cities in Jiangsu Province as the research unit. The first-level fuzzy comprehensive evaluation method was adopted to measure the compactness of the traffic flow, investment flow, and innovation flow networks, based on which the spatial correlation map of the network compactness was constructed. Taking these as the explained variables, the exponential random graph model was introduced to form an estimation model from four aspects (including network endogenous mechanism, optimization mechanism, proximity mechanism, and coordination mechanism) to quantitatively analyze the formation mechanism of the urban network compactness in Jiangsu Province. The study revealed that the network compactness of Jiangsu Province is of a significant closed triangular structure in which local clustering is conspicuous, and there is relatively significant local compactness, while the improvement of the compactness level does not depend on the organization or coordination of third-party cities. Network compactness demonstrates a feature of powerful scale orientation, where cities with higher economic scale tend to build compact relationships. Network compactness comes into formation under the action of multi-city spatial clustering. The complex spatial relationship arising from geographical proximity is the premise of formation of the network compactness. However, the factor flow restrictions incurred by regional division in the province have restricted the establishment of the compactness relationship among cities, and then the network compactness at the region level demonstrates the phenomenon of abnormal clustering.

Key words: urban network, compactness, formation mechanism, exponential random graph model, Jiangsu Province