Forecasting passenger load for a fixed planning horizon

Abstract

This paper presents a forecasting system for predicting the number of boarding passengers on each of the next N departure dates of a particular flight leg using as input the booking levels made for the next N departure dates of that flight leg. A forecasting system that gives good estimates of future boardings of a particular flight leg is an important and essential tool for route management in the airline industry. The proposed forecasting system consists of a Kalman filter, which optimally integrates a large class of reservation measurements from a variety of sources and at arbitrary times, which is the case of the data, considered here. The accuracy of the predictions for a case study proved to be promising.