A competitive approach to airline revenue management

Abstract

Since the 1980s, the airline industry has seen two major changes: Deregulation, which led to an increase in competition, and the development of revenue management systems. Paradoxically the revenue management models used have not incorporated many competitive considerations. In this thesis we study the interactions between existing airline revenue management systems in competitive markets. We use the Passenger Origin Destination Simulator (PODS) for simulations. After a review of the past research on such interactions, we develop our own model and use simulations to identify and measure the extent to which revenue management systems of competing airlines affect each other. The model introduced highlights the importance of spill-of-demand between airlines. We show that with current revenue management practice, a legacy carrier should be less sensitive than a low-cost carrier to revenue management competitive interactions. As compared to an equivalent monopoly, an airline oligopoly tends to allocate more seats to high-fare passengers and fewer seats to low-fare passengers. With steady demand distributions, an airline's expected revenues are a decreasing function of the seat capacity allocated by its competitors to high-fare passengers. Existing revenue management systems react to competitor moves automatically only if a change in the seat allocation rule by an airline occurs over a large enough number of successive departures to be detected by forecasters. We then suggest how to improve revenue management based on the interactions identified. With steady demands, if a competitor increases (respectively decreases) its seat allocation for high-fare passengers, the best response to optimize revenues on the short-term is to decrease (respectively increase) seat allocation for high-fare passengers. We also show that the use of EMSRb-optimization by competitors results in a near-optimum competitive equilibrium, and a near-optimum cooperative equilibrium if airlines do not share revenues. Rarely can competitive interactions justify an airline to override the EMSRb seat-allocation rule to optimize revenues. Last, we introduce LOCO-based Forecast Multiplication, a heuristic forecast adjustment made in response to the current seat availability of the competitors that can increase an airline's revenues substantially.