Probabilistic Airspace Congestion Management

It is difficult to accurately predict air traffic conditions, such as the traffic demand on a particular Air Traffic Control sector, hours in advance. Convective weather, when present, reduces sector capacity and is also difficult to predict. Nonetheless, current procedures and decision-support systems for strategic traffic management do not explicitly acknowledge the uncertainty in traffic and weather predictions. The result is highly-conservative decision-making, producing unnecessary delays and underutilized airspace.

This paper presents a novel approach to airspace congestion management, in which prediction uncertainties are measured and explicitly applied to improve decision-making. A probabilistic congestion forecast will focus the attention of decision makers on the most likely problem areas. Automation can use this forecast to select incremental congestion resolution actions to manage the risk of future congestion while deferring potentially unnecessary impacts and delays whenever possible. A probabilistic congestion forecast can also be used to estimate the risks for each flight, and this can be provided as feedback to system users (e.g., airlines) before resolution actions are taken, so that users can take preventive action, or not, based on their own priorities and tolerance for schedule risk. This approach has the potential for reducing delays and other impacts while maintaining system safety and increasing the opportunities for users to make decisions about their own flights.