Peter BELOBABA

Theory Versus Revenue Management Realities

Operations research models applied to the airline revenue management problem have relied heavily on assumptions concerning the ability of the airline to accurately forecast demand at a level of detail corresponding to the assumptions of the optimization algorithms and, then, to implement the solutions of the optimization algorithms by effectively controlling
passenger requests and seat inventories. The reality of airline revenue
management in practice is that the underlying characteristics of passenger
demand and the ability of airlines to effectively control bookings simply do not match the assumptions of these optimization models.

In this presentation, we examine the impacts on airline revenue management performance of the discrepancies between most RM optimization models and reality, by simulating realistic scenarios with the Passenger Origin-Destination Simulator (PODS). First, we illustrate how the requirement for accurate independent demand forecasts as inputs to RM optimizers leads to situations under more realistic simulated conditions in which the “best” forecasts that maximize airline revenues are in fact not the most “accurate”. Second, we demonstrate how current limitations of airline reservations systems constrain the ability of airlines to actually implement their network “optimal” RM solutions, negatively affecting RM revenue gains.