Overbooking models for air cargo yield management

Abstract

This thesis addresses cargo-overbooking problem in the airline industry. A mathematical model is formulated to determine the optimal overbooking level with minimum expected under-sales cost and spoilage cost. The cargo booking process is modeled as a two-dimension Markov process. It is a combination of homogeneous Poisson arrival process with a constant arrival rate and non-homogeneous Poisson cancellation process with a cancellation rate depending on the number of customers in the system. The two state variables are defined as the number of bookings in the system and the total amount of space that has been booked. Limiting probability distribution is used to approximate the joint probability distribution of the final number of bookings and final amount of cargo coming for boarding. Simulation results show that the overbooking model produces very good approximations on real booking behaviors and ensures the minimum expected total cost in the assumed setting.