Assessment of Node Importance in Air Transportation Networks using Multi-Criteria Decision Analysis

Complex network theory provides powerful tools to analyze and understand the structures and dynamics of complex air transportation systems. One example for a successful application of complex network theory in air transportation systems are airport networks, where airports are nodes, and two airports are connected if there exist flight connections. In complex networks, a node can be characterized by several centrality metrics: Degree, betweenness, closeness, eigenvector, etc. How to identify which node is most important or most critical, regarding different notions of centrality, is a typical multi-criteria decision problem. This paper incorporates Multiple Criteria Decision Analysis (MCDA) techniques to assess node importance in air transportation networks. First, we investigate the correlation among a large set of metrics in order to identify independent, representative metrics. Second, we derive preferences among multiple metrics using Entropy method. Third, we apply an improved MCDA method which can maintain the ranking consistency of the alternatives. In the end, we perform uncertainty assessment for preferences among multiple metrics. Our proposed methodology is applied to an airline network.