How China Southern Airlines can use Graph Theory to improve its reliability
While researching the profound theories of networks and graphs is crucial and valuable, their applications make the field especially fascinating. In fact, the applications of graphs and networks are incredibly diverse and significant. In this case, a research team used graph theory and networks to help Chinese Southern Airlines (CSA), a local Chinese airline company, fix a critical operations issue and provide recommendations to elevate CSA’s market position towards becoming an industry leader.
The research paper attacks one of CSA’s core problems: reliability. As a local airline, CSA has intense competition with larger airlines that have extensive regional and international networks. However, if CSA can emerge as the obvious choice for local travel, it can thrive as the strongest regional airline in China. To do so, the airline has to win customer loyalty through reliability and access. The research paper investigates how CSA can achieve that through applying complex network theory and analysis.
To construct a working model of CSA’s flight network, the research team begins by establishing what the nodes and edges represent. Each node indicates an airport in the network, and each edge denotes a flight that connects from one airport to another. Because the data did not provide enough information on the direction of the flights, the network is unidirectional, where edges only indicate that a flight connects the two airports. To accurately assess and evaluate CSA’s network reliability, a strong methodology includes looking at local bridges, average path length, and strength of ties. Local bridges indicate the airports that have no other airport in common. While it’s still possible to get from one airport to another, the path length would be incredibly lengthy and not ideal for customers. Average path length indicates how connected the graph is by assessing how long it takes to get from a given airport to another, on average. The strength of the tie signifies the frequency of flights between the two airports. If flights are not frequent enough for customers, the reliability of the airline will diminish.
Given the working model of CSA’s flight network, the research paper conducts extensive empirical analyses to determine key data points of CSA’s reliability. For example, the research paper identifies the CSA network to have 187 nodes and 1245 edges. On top of that, the average degree of the nodes is 6.68, indicating that, on average, each airport is directly connected to 6 other airports. The average path length between the nodes is 2.558, meaning the average shortest length from a given airport to another is about 2.5 trips. The clustering coefficient of the network is 0.592. Given the small average shortest path length and large clustering coefficient, the research paper notes the CSA network belonged to a small-world network, meaning that most nodes are not directly connected to many other nodes, but it is quite easy to get from any node to another without a long path. It also found the “betweenness” values, which indicate the influence of that node as a bridge between other node pairs.
Putting all the statistical data points together, the research paper identifies a few key points. Unfortunately, some airports tend to be critical elements to the network, and many routes rely on their contributions as a connection. For example, the Guangzhou airport has a betweenness value of 51.414, double that of the second-highest betweenness value, and a degree of 106, which is more than double that of the second-highest degree value. If this critical airport were to shut down temporarily due to bad weather, for example, the impacts on the network and average path length would be detrimental. Not just that, the research team concluded that if just two nodes were removed, the network would essentially break down. Clearly, the data points tell a compelling narrative that CSA has serious reliability challenges that it should remedy to excel its market position amongst larger airlines.
The research team also outlines a few recommendations for CSA. Firstly, CSA should ensure that the Guangzhou airport stays efficient and well-managed. If the airport were to shut down temporarily for any reason, the CSA network would become severely diminished. At the same time, CSA should perform a deeper analysis of the edges to increase the clustering coefficient, lower the average shortest path length, and strategically boost the weight of the ties (increase flight frequency) that represent the busiest and most significant legs to the flight network.
Using graph theory and its applications on the CSA business model, travel logistics, and operations, CSA has an exciting opportunity to become a market leader within the Chinese domestic airline travel industry. By doing so, CSA will be able to elevate its reliability and capture meaningful customer loyalty. CSA can then seize opportunities to dominate the local Chinese airline market and become the go-to airline for domestic airline travel in China.
Author: Ethan Talreja
Source: https://www.scielo.br/j/jatm/a/KY3yXWmFcbCTRrgS4VjYZjm/?format=pdf&lang=en