Airline Network Analysis And Airline Scheduling System Design

The airline network is a constructed system consisting of flight legs. The airline networks have a strong influence upon the profitability, operational efficiency, and quality of service of the airlines. The developed world has been studying the airline network for over twenty years, yet the academic world of China still mainly focuses upon qualitative analysis, in lack of deeper qualitative analysis. Recently, new discoveries in the area of complex networks– the small-world effect and scale-free property arouse a new wave of study interest in different disciplines.In this thesis, we study the airlines through theories in complex networks. Our study mainly focuses upon two perspectives. First of all, the network topological properties. We study the airline topologies of 17 major airlines at home by computing their edges, degree distribution, betweeness distribution, clustering coefficient, and network diameter. We discover that most of Chinese domestic airline networks are small-world scale-free networks. Secondly, the network robustness. We compute the robustness value of different airline works, and compare them with each other. Our conclusion is that most small airlines at home have very good robustness. In addition, in order to improve the network fitness, we optimize the current network topology by minimizing the network diameter and maximizing the passenger flow on each edge.After dozens of years' development, domestic airlines could meet customers' needs in terms of revenue management and operational management. However, the areas of airline network design and fleet management are still underdeveloped. Chinese airline networks are experiencing the evolution from point-to-point networks to hub-and-spoke networks. This evolution, nevertheless, is in lack of theoretical support. We therefore forward a mathematical which combines the airline network design and fleet assignment, the aim of which is to minimize the operational cost. Also we forward a two-stage simulated annealing algorithm to compute it. Our experiment upon the data of passenger flows among Chinese cities proves that the efficiency of our algorithm is superior to the classical SA algorithms forwarded by Ernst.