Research On Robustness Of Air Transportation Network And Flight Delay Correlation

The understanding of the structure and dynamics of the multilayer air transportation system is of great important to both academic field and industry world.Based on the empirical datasets,this thesis investigates the structure and dynamics of the air transportation system by the leverage of network science and data science while taking into the operational characteristics of the system.More specifically,this thesis analyzes the structure,robustness,and flight delay correlations of the air transportation system.Firstly,a multilayer air transport network is proposed from recently emerged multilayer network science.The air transport network consists of airport network,air route network and sector network.The concept of sector network is for the first time proposed which is based on the system operation.The characteristics of each network is thoughtfully investigated.Secondly,the robustness of air transport network is studied from two perspectives,namely network structure and operational dynamics.From network structure perspective,we attack the airports by removing them from network with different strategies.From operational dynamics perspective,we define strategies which would cause the largest losses in network connectivity to detect critical nodes.The traditional robustness studies by removing the nodes from network to simulate the destructive impact simply,but the dynamics of system is neglected.We investigate air transportation system resilience when the disruptive events occur through the reduction of airport capacity.Different attack scenarios are simulated.We found that the system exhibits a process as stabilize – deteriorate – recover – stabilize while under attack.The results indicate that structural indicators can roughly distinguish the importance of nodes,but they are unable to identify critical nodes when considering operational characteristics.Finally,this thesis investigates spatial correlation of flight delay,and finds that delay has some relevance to the distance between the airports.The geographically close airports are more correlated.The findings in this thesis have significant meanings to both academic field and industry world.