Passenger overbooking, using no-show probability

Abstract

Passengers that fail to show up for the flights they have booked are called no-show passengers. No-shows are an inevitable part of aviation and as a reaction to no-shows, airlines often overbook flights to optimize their resources. Overbooking can lead to some passengers being denied to board a flight if the bookings exceed the seating capacity. That can be costly not only because of possible penalties and cost of rerouting, but also due to the harm it can have on the reputation and image of the airline. This research analyzes what variables affect the number of no-show passengers through data analysis and interviews. It was discovered that the probability of no-show passengers on board flights follows a geometric distribution. By conducting a regression on the probability of no-show passengers, the significant variables explaining the probability were analyzed as well. A revenue model was created by using the no-show probability to find the optimal overbooking level when the probability of denied boarding is taken into consideration. In total six scenarios were computed and simulated for different probabilities of all passengers showing up and cost ratios due to denied boarding.