Abstract Essential to effective revenue management decisions in the airline industry are good forecasts of future passenger demand for tickets. Since 1986, BehavHeuristics, Inc. (BHI) has pioneered the application of neural networks as a forecasting method for the airline problem. Our research and implementation have shown that not only are neural networks a viable alternative to more traditional methods of forecasting, but that significant improvements in accuracy can be achieved. In addition we have found that their flexibility and ease of use make them ideal as the basis for a revenue management system which can be quickly adapted to different airlines. In this paper we describe the general problem of forecasting airline demand and discuss a variety of issues related to the system application. Comparisons of this technique to other traditional techniques are then made on actual airline data to show that improvements in forecast error can be significant.
A Large-Scale Neural Network for Airline Forecasting in Revenue Management
1998-01-01
22 pages
Article/Chapter (Book)
Electronic Resource
English
Forecasting for airline network revenue management : revenue and competitive impacts
DSpace@MIT | 1998
|DSpace@MIT | 2016
|