Optimizing flights of a fleet of aircraft using a reinforcement learning model

1. An apparatus for optimizing flights of a fleet of aircraft, the apparatus comprising:

a memory configured to store computer-readable program code; and

processing circuitry configured to access the memory, and execute the computer-readable program code to cause the apparatus to at least:

access flight plans for flights of a fleet of aircraft through an air transportation network, each flight subject to compliance with air traffic control requirements;

track respective positions and respective trajectories of one or more aircraft of the fleet of aircraft that are currently in-flight;

apply the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight to a reinforcement learning model configured to determine maneuvers starting at the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are in-flight or starting at respective beginnings of the flights of the fleet of aircraft that are not in-flight that achieves a respective maximum cumulative value of an operational efficiency metric across the flights of the fleet of aircraft, one or more of the maneuvers constituting a deviation from a respective flight plan that propagates to other aircraft of the fleet of aircraft that are caused to perform responsive maneuvers to maintain compliance with the air traffic control requirements;

perform a comparison of respective maximum cumulative values of the operational efficiency metric for the aircraft of the fleet of aircraft to determine a ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values;

select one of the aircraft based on the ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values; and

send a notification of the deviation from the respective flight plan to the one of the aircraft.

2. The apparatus of claim 1 , wherein the processing circuitry is configured to execute the computer-readable program code to cause the apparatus to further at least:

generate simulated flight plans for simulated flights of the fleet of aircraft; and

train the reinforcement learning model using the simulated flight plans as a training dataset.

3. The apparatus of claim 1 , wherein the flight plans indicate flight paths of the aircraft of the fleet of aircraft, and the deviation is from one or more of a planned altitude, speed or heading of the aircraft on a respective flight path.

4. The apparatus of claim 1 , wherein the apparatus caused to send the notification includes the apparatus caused to send the notification that also includes an explanation for selecting the one of the aircraft.

5. The apparatus of claim 1 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, and the one of the aircraft is selected, prior to at least a respective flight of the one of the aircraft, and

wherein the processing circuitry is configured to execute the computer-readable program code to cause the apparatus to further generate an updated flight plan for the one of the aircraft that includes the deviation from the respective flight plan, and the apparatus caused to send the notification includes the apparatus caused to send the updated flight plan to the one of the aircraft.

6. The apparatus of claim 1 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, the one of the aircraft is selected, and the notification of the deviation is sent to the one of the aircraft, during at least the respective flight of the one of the aircraft.

7. The apparatus of claim 1 , wherein the operational efficiency metric includes one or more of fuel consumption, carbon emissions or flight time across the fleet of aircraft.

8. The apparatus of claim 1 , wherein the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight are repeatedly applied to the reinforcement learning model throughout the flights of the one or more aircraft to determine updated maneuvers that achieve the respective maximum cumulative values of the operational efficiency metric across the flights of the fleet of aircraft.

9. A method of optimizing flights of a fleet of aircraft, the method comprising:

accessing flight plans for flights of a fleet of aircraft through an air transportation network, each flight subject to compliance with air traffic control requirements;

tracking respective positions and respective trajectories of one or more aircraft of the fleet of aircraft that are currently in-flight;

applying the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight to a reinforcement learning model configured to determine maneuvers starting at the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are in-flight or starting at respective beginnings of the flights of the fleet of aircraft that are not in-flight that achieves a respective maximum cumulative value of an operational efficiency metric across the flights of the fleet of aircraft, one or more of the maneuvers constituting a deviation from a respective flight plan that propagates to other aircraft of the fleet of aircraft that are caused to perform responsive maneuvers to maintain compliance with the air traffic control requirements;

performing a comparison of respective maximum cumulative values of the operational efficiency metric for the aircraft of the fleet of aircraft to determine a ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values;

selecting one of the aircraft based on the ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values; and

sending a notification of the deviation from the respective flight plan to the one of the aircraft.

10. The method of claim 9 further comprising:

generating simulated flight plans for simulated flights of the fleet of aircraft; and

training the reinforcement learning model using the simulated flight plans as a training dataset.

11. The method of claim 9 , wherein the flight plans indicate flight paths of the aircraft of the fleet of aircraft, and the deviation is from one or more of a planned altitude, speed or heading of the aircraft on a respective flight path.

12. The method of claim 9 , wherein sending the notification includes sending the notification that also includes an explanation for selecting the one of the aircraft.

13. The method of claim 9 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, and the one of the aircraft is selected, prior to at least a respective flight of the one of the aircraft, and

wherein the method further comprises generating an updated flight plan for the one of the aircraft that includes the deviation from the respective flight plan, and sending the notification includes sending the updated flight plan to the one of the aircraft.

14. The method of claim 9 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, the one of the aircraft is selected, and the notification of the deviation is sent to the one of the aircraft, during at least the respective flight of the one of the aircraft.

15. The method of claim 9 , wherein the operational efficiency metric includes one or more of fuel consumption, carbon emissions or flight time across the fleet of aircraft.

16. The method of claim 9 , wherein the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight are repeatedly applied to the reinforcement learning model throughout the flights of the one or more aircraft to determine updated maneuvers that achieve the respective maximum cumulative values of the operational efficiency metric across the flights of the fleet of aircraft.

17. A computer-readable storage medium for optimizing flights of a fleet of aircraft, the computer-readable storage medium being non-transitory and having computer-readable program code stored therein that, in response to execution by processing circuitry, causes an apparatus to at least:

access flight plans for flights of a fleet of aircraft through an air transportation network, each flight subject to compliance with air traffic control requirements;

track respective positions and respective trajectories of one or more aircraft of the fleet of aircraft that are currently in-flight;

apply the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight to a reinforcement learning model configured to determine maneuvers starting at the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are in-flight or starting at respective beginnings of the flights of the fleet of aircraft that are not in-flight that achieves a respective maximum cumulative value of an operational efficiency metric across the flights of the fleet of aircraft, one or more of the maneuvers constituting a deviation from a respective flight plan that propagates to other aircraft of the fleet of aircraft that are caused to perform responsive maneuvers to maintain compliance with the air traffic control requirements;

perform a comparison of respective maximum cumulative values of the operational efficiency metric for the aircraft of the fleet of aircraft to determine a ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values;

select one of the aircraft based on the ranking of the aircraft of the fleet of aircraft according to the respective maximum cumulative values; and

send a notification of the deviation from the respective flight plan to the one of the aircraft.

18. The computer-readable storage medium of claim 17 , wherein the computer-readable storage medium has further computer-readable program code stored therein that, in response to execution by the processing circuitry, causes the apparatus to further at least:

generate simulated flight plans for simulated flights of the fleet of aircraft; and

train the reinforcement learning model using the simulated flight plans as a training dataset.

19. The computer-readable storage medium of claim 17 , wherein the flight plans indicate flight paths of the aircraft of the fleet of aircraft, and the deviation is from one or more of a planned altitude, speed or heading of the aircraft on a respective flight path.

20. The computer-readable storage medium of claim 17 , wherein the apparatus caused to send the notification includes the apparatus caused to send the notification that also includes an explanation for selecting the one of the aircraft.

21. The computer-readable storage medium of claim 17 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, and the one of the aircraft is selected, prior to at least a respective flight of the one of the aircraft, and

wherein the computer-readable storage medium has further computer-readable program code stored therein that, in response to execution by the processing circuitry, causes the apparatus to further generate an updated flight plan for the one of the aircraft that includes the deviation from the respective flight plan, and the apparatus caused to send the notification includes the apparatus caused to send the updated flight plan to the one of the aircraft.

22. The computer-readable storage medium of claim 17 , wherein the flight plans are accessed and applied to the reinforcement learning model, the comparison is performed, the one of the aircraft is selected, and the notification of the deviation is sent to the one of the aircraft, during at least the respective flight of the one of the aircraft.

23. The computer-readable storage medium of claim 17 , wherein the operational efficiency metric includes one or more of fuel consumption, carbon emissions or flight time across the fleet of aircraft.

24. The computer-readable storage medium of claim 17 , wherein the flight plans and the respective positions and respective trajectories of the one or more aircraft of the fleet of aircraft that are currently in-flight are repeatedly applied to the reinforcement learning model throughout the flights of the one or more aircraft to determine updated maneuvers that achieve the respective maximum cumulative values of the operational efficiency metric across the flights of the fleet of aircraft.