Air Route Network Generation And Optimization Based On Compelx Network Theory

With the rapid development of civil aviation transportation enterprises,the air traffic flow continues to growth,thus,the airspace congestion,traffic control,flight delays and a series of problems were happening with increasing frequency.As a carrier of air transportation,air route network is the physical space to realize air transportation operations,its structure is the important factors that has a huge effect on the safety,efficiency and economic of air traffic flow.Therefore,the generation and optimization of air route network becomes an urgent subject which need to be studied as an airspace planning technology in the field of air transportation.Firstly,paper elaborated the research status of air transportation network planning and introduces the basic complex network theory as well as method of network planning.From the perspective of the development and evolution of the network,using weighted link prediction algorithm to optimize and predict airport network structure,and determine the network edge similarity index;Using the K-Core algorithm to implement the layering of airport nodes and apply community partition algorithm which based on the closeness centrality to realize community divided;Through the optimal path search algorithm based on preference and the K shortest path search algorithm based on constraint,initial air route network was generated.Then,realizing the clustering of cross nodes by improved the DBSCAN clustering algorithm.With the total operation cost minimum as the objective,the range of mobile and average network nonlinear coefficient as constraints,the network nodes layout optimization model is established and solved by the PSO algorithm and local search strategy.Finally,taking the 200 commonly used airports in mainland China as an example,paper analyzed its corresponding airport network and air route network,also,optimized and predicted the airport network structure which carried out 24 addition airlines and the similarity index of every airline;Airports were divided into three layers,namely the Core layer,Bridge layer and Periphery layer,and airports in Bridge layer were divided into 47 communities;There were 804 cross nodes in the initial generated air route network,after DBSCAN clustering,its number decreased by 547.The air route network finally generated after the cross nodes layout optimization.In the final network,the number of cross nodes is 242 and the average network nonlinear coefficient is 1.0991,which strongly verified the feasibility of proposed scheme about air route network generation and optimization.