Design and implementation of a sector-based airspace model for the MIT Extensible Air Network Simulation

Abstract

The MIT Extensible Air Network Simulation (MEANS) is a tool that has been designed to assist airline schedulers and air traffic managers in predicting flight delays for given air traffic scenarios. One aspect of the simulation, the determination of flight times, has received criticism from the MEANS users as being too simplistic for their needs. Currently, MEANS predicts flight times based on a historical distribution of observed flight times between city pairs. This system ignores the effects of flight level winds and airspace congestion, two major determiners of flight time. The replacement flight time model presented divides the airspace into discrete sectors based on existing divisions in air traffic control. Each sector has its own wind conditions and capacity limitations which affect passing flights. Results show that, after some calibration, the new flight time model produces accurate flight times when the airspace is divided into ARTCC domains and does not introduce additional errors into other parts of the simulation. Additionally, test scenarios show that the new system is capable of modeling airspace capacity events, such as a radar failure. Comparative results reveal that the old, distribution model produces surprisingly accurate flight times for typical wind conditions and airspace utilization.