Modeling airline group passenger demand for revenue optimization

Abstract

Many airlines currently use a variety of analytic techniques for seat inventory control as part of a larger revenue or "yield" management system. However, much of the this effort has emphasized decisions based on the individual passenger, and has neglected a significant segment of total airline passenger demand, namely that of group passengers. Group passenger demand differs in several important respects from individual passenger demand, and these differences have motivated the need for separate attention in booking procedures and future demand forecasting. In this thesis we begin by discussing the issues involved with trying to characterize the stochastic nature of airline group passenger demand, and identify the primary elements of variability associated with it. Later, we use these primary elements of demand to develop a mathematical model for the distribution of group passengers on a given flight(s). Armed with a well-defined distribution for group passenger demand, we enhance current mathematical programming approaches for solving the seat inventory control problem to include the control of group seat inventories. We then present a model for determining the minimum per passenger fare an airline should charge an ad hoc group request based on the displacement of individual passengers. Finally, we discuss the issues involved with overbooking in the group demand context, and suggest areas for further research.