Relative trajectory cost estimation for CTOP applications using multivariate nonparametric finite mixture logit

Abstract

We study airline decision-making in response to the Federal Aviation Administration’s Airspace Flow Program (AFP) to empirically describe policy-relevant behavioral strategies and preference structure of air carriers. Using observed responses made by airlines, we infer utility functions of different route options in AFP with respect to flight time and arrival delay using a finite mixture latent class choice model. We empirically describe the trade-off that airlines face between flight time and departure or arrival delay. The average estimated cost of airborne delay is 24 times greater than the cost of ground delay. Airborne cost per minute is over 100 greater than ground delay cost for large carriers and 4–20 times greater for small airlines. The results are applicable to future traffic flow management programs, such as Collaborative Trajectory Options Program (CTOP).