机场体系中心性的网络分析方法与实证

Network Analysis on Centrality of Airport System

机场体系是航空运输组织网络的核心组成部分,认知其空间结构特征对航空运输网络的建构及辨析其表征的城镇体系结构等具有重要的意义,而中心性在刻画机场体系空间结构特征中具有基础性的作用。在研究Christaller中心地理论及其分析方法的基础上,引入网络分析方法的度中心性、邻近中心性和介中心性,这3个指标分别反映了节点在网络中的直接交流能力、可达性服务水平和控制网络交流的能力。以中国航空网络为例,研究表明网络中心性指标对机场体系的空间结构具有较好的解释性。度中心性和邻近中心性表现出较好的一致性,但与介中心性差异较大。尽管北京和上海的各类中心值始终位居前两位,但对于度中心性和邻近中心性2个指标,第三、四和五位分别为广州、深圳和成都;而对于介中心性,取而代之的是昆明、广州和乌鲁木齐,更多地体现了地缘政治的影响。综合3类网络中心性测度方法和中心地理论,提出具有探索性意义的系统中心性分析方法,其研究结果较好地体现了中国机场体系的等级结构、上层系统"鼎形"集聚特征和国家地域系统的"鞍"型效果,且与《全国民用机场布局规划》的五大机场体系具有较好的耦合性。

The airport system is an important component for the organization of aviation transportation. It is an effective way to analyze the constitution of air transportation and urban systems by identifying the spatial structure of airport system. Centrality is one of the basic methods. Based on network analysis, three common indices, degree centrality, closeness centrality and betweenness centrality were used to measure centrality for individual cities in the paper. That is to say, ＂Being central＂ is not limited to being connected to others, but also being close to all others and being intermediary between others. In the airport system of China, degree centrality shows the ＂Ding Pattern＂ for the top class and the ＂Flyover Effect＂ for the national network. Closeness centrality makes it measurable for the airport service. Betweenness centrality is a good measure to analyze regional hubs and the core-periphery pattern. On individual cities, Beijing and Shanghai are clearly the top two central cities in all three indices; and rankings on other cities by the betweenness centrality differ significantly from those by the degree and closeness centralities. Based on the three centrality indices, an Exploratory Weighted Method （EWM） were put forward, namely system centrality, to do a further research. The results showed the hierarchical structure with the ＂Ding Pattern＂ in the top class and the ＂Flyover Effect＂ in the national network. Besides, the spatial pattern by the system centrality accords well with the five airport clusters designed in the National Civil Airports Deployment Plan.