KOTI - Korea Transport institute

The airport is arguably one of the most complex systems in existing transportation facilities in that it is operated by a diverse group in addition to stakeholders and is influenced heavily by external factors. For example, in the event of a plane accident, facility incident or other airport-related incident, the number of responsible stakeholders involved is greater than any other transportation sector. Despite the importance of robust maintenance of the facility, in recent years, we are observing signs of an increase in potential airport operation failure due to the rapid increase in aviation demand, growing expectations of aviation services, and an increase in automation and connection systems with the introduction of advanced systems.
The Republic of Korea is not the only country that is experiencing these issues. The United States of America (USA) and the United Kingdom (UK) has also experienced these issues in the past. In order to resolve such problems, they have shifted away from responding to the emergency situations and they have introduced airport irregular operations when responding to disruptions of normal flight schedules. Such efforts are also being made in Korea to improve the efficiency of airport operations and implementation of management systems to prevent and respond to serious accidents. However, there are no specific definitions or methodology to determine airport irregular operations and ripple effects for predictive and preemptive responses as detailed studies have not been conducted. In order to solve these problems, the following study took an incremental approach to newly define airport irregular operations by distinguishing between terms for emergency situations and normal situations. Furthermore, the study suggests a methodology for identifying the concept of the matrix based on severity and duration rather than cause. An index for the airport operating system was set up as an increase in delay time and a tool was developed to simulate a delay situation based on actual operation data from individual airports. We conducted simulations for airports located in Jeju, Gimpo, and Gimhae Airport where air traffic congestion is greater than other airports in Korea to establish the classification of airport irregular operation based on unsupervised learning of the results.
A scenario analysis was carried out to verify the constructed methodology assuming that the airport irregular operations occur. In addition, we have verified characteristics of constructed model through ensemble learning, which is a type of supervised learning method, and confirmed its applicability. It was difficult to apply constructed classification model directly due to the data limitation but the study suggested potential applicability in the future once we accumulate more data and refine methodology.
For policy utilization, we examined current airport irregular operation response systems in Korea and conducted a survey for passengers related to flight delays and cancellations. The result suggested that policy consideration is needed to improve the response system rather than focusing on the improvement of facilities. It is also imperative to gain more information to satisfy customers and passengers in the airport. The developed methodology in this study can be utilized as a tool to meet these policy needs.
The study suggests alternatives to emergency response systems with a manual, and in addition, the need for maximizing the effectiveness of the response systems through the introduction of the BCP concept and a specific plan for the group decision making process. Moreover, the study briefly reviewed the quantitative expectation effect through preemptive response to the airport irregular operations and analyzed the social benefits.
Finally, the study comprehensively covers development of basic analytical methodology based on the policy needs of stakeholders in airport operation, and its application to policies such as system and manual improvement. The approach suggested in this study is different to the existing approach, so it is necessary to enhance publicity and a bond of sympathy, implement the learning method and develop a manual for airport irregular operations through actual verification and data accumulation. We expect airport operations to be seamless through preemptive and predictive response as described above.