当下城市网络研究中存在相对重结构特征测度、轻影响机理分析的问题,而在城市网络影响机理的研究中,又存在重理论定性的宏观描述或基于独立变量的常规性统计分析,轻从图论、结构视角的城市网络微观发育机理研究。因此,从结构关系依赖视角出发,基于苹果手机2019年供应商数据构建研发型、生产型、代工服务型全球化城市网络,运用指数随机图模型的微观构型测度城市网络的生长发育机理。结果表明：（1） 互惠性与中介效应在3类城市网络普遍存在,深刻影响城市网络的通性发育机理。（2） 偏好依附过程与接收者、发送者效应是解释城市网络等级中心性发育的结构机理,二者都体现了城市网络发育中以入度为核心的路径依赖现象。（3） 三角形结构（传递三角形与循环三方组）和同质性是促进城市网络集群发育、富人俱乐部现象的微观基础,同质性对城市网络的发育机理主要体现在核心城市间的联结关系上。（4） 企业路径依赖与距离是影响城市网络发育的核心外生因素。其中,对距离的分类讨论发现,地理邻近性对3类城市网络存在便在性影响。认知邻近性在研发型城市网络对长距离联系的影响与敏感度更大;在生产型城市网络中对中等距离有正向影响,代工服务型城市网络对短距离阈值敏感。该研究对丰富与拓展现有城市网络影响机理的研究视角有重要意义。

Recently there has been an increase in the number of the relationship turns and the evolution of new economic geography, which facilities the transformation of city network research, namely, from structural characteristics to influence mechanism. In light of the existing research progress, the paper introduces microconfiguration through the lens of the structural perspective to investigate the growth and development mechanism of city networks. To be more specific, initially, the city networks of R&D-oriented, production-oriented, and OEM service-oriented are specified built based on the suppliers’ data of iPhone’s components and parts in 2019, and afterward, we conduct motif analysis and exponential random graph model (ERGM) to examine the microprocesses in the spatial growth of city networks. The results demonstrate that: (1) reciprocity and the intermediary effect are widely and commonly distributed in the three types of city networks, which profoundly affect connectivity development and promote the growth of the hierarchical structure and the availability of the entire network. (2) The process of preferential attachment and the effect of receiver and sender (branches) constitute core structural mechanisms interpreting hierarchical centrality pattern of city networks, both of them embody the phenomenon of path dependence centered on in-degree centrality in the growth processes of city networks. (3) Triangle structure (transitive triangle and circular triad motif) and homogeneity are micro basis components that promote the cluster growth and community structure of city network and the phenomenon of rich clubs. Moreover, the homogeneity influence on the spatial growth of city networks is primarily reflected in the linkages between core cities. (4) Enterprises’ path dependence and distance are the main external factors that drive and affect the growth of city networks. Overall, in the current city network research, there is still a problem that quite focuses on measuring structural characteristics, while ignoring the analysis of the influence mechanism. Furthermore, in the existing literature on the influence mechanism of city networks, many researchers focus on a qualitative macro description based on independent variables, nevertheless, but little on the micro growth and development mechanism of city networks from the standpoint of graph theory and structuralism. Hence, a microconfiguration of ERGM is used through the lens of relationship dependence in this paper, subsequently; we introduce exogenous covariates and endogenous structural variables into the hypothesis verification of city network influence mechanism, which is very important for broadening the research scope of the influence mechanism of the existing city network research.