Challenge
The operation of an airline or a complete airport is a complex and multidimensional problem. In the operations center of an airline, optimal processes are worked on at high pressure every day. The punctual departure of an aircraft can only be guaranteed if many different factors are met, such as ideal weather, an appropriate traffic situation in the airspace or sufficient reserve capacity. As a result, there is often little time to proactively address anticipated problems in the flight schedule in the days ahead. Our customer faced the challenge of wanting to support the operations center by using a predictive planning tool.
Approach
To support the operations center, we developed an application that uses machine learning models to predict which risk factors will be relevant for planning in the coming days. The application serves as an information hub for network planning for future operation days. It minimizes the manual effort required to gather information by automatically compiling relevant information and making it available in a form that can be individually structured. Furthermore, the application shows predictive risk estimates for various factors such as crew reserves, ATC delays or ground exchange possibilities for the next four days.
Results
Based on the daily forecasts and comprehensive information from the application, we were able to help our client establish and improve ways to proactively act on future flight plans. The implemented machine learning models improved the precision of risk assessments to enable optimal flight plan adjustments for future days.