Revenue management under demand driven dispatch

Abstract

Demand Driven Dispatch is an operational mode by which airlines can dynamically reassign aircraft types of the same family but of differing capacities to better match capacity with demand in a profit maximizing way. Current algorithms for dynamic reassignment of aircraft types suggest that the nect assignment problem be solved repeatedly during the booking horizon of the pool of flights being considered in the type assignment model. This approach has been shown to produce a 1% to 5% improvement in operating profits when compared to the static fleet assignment currently practiced by airlines. In our work we formalize and explore the necessary interaction between a revenue management model and the fleet assignment decisions that the airline must make under the Demand Driven Dispatch operational mode. Based on our findings. we produce a set of alternate algorithms for dynamic fleet type reassignment which significantly reduces the number of fleet assignment problems that must be solved during the booking horizon of the pool of flights being considered. Using demand data from a major U.S. airline, we simulate both the traditional algorithms used for dynamic fleet reassignment and the alternate algorithms developed in this thesis. Our results show that it is indeed possible to delay the first fleet assignment decision that the airline must make, thus reducing the number of fleet assignment problems that must be solved. Further, we show that our approach can even outperforms the traditional Demand Driven Dispatch algorithms both in terms of revenues and in terms of passenger loads, by integrating the delayed fleet assignment decision with the revenue management process.