Adaptive Mechanisms in an Airline Ticket Demand Forecasting System.

Abstract

Adaptivity is a very important feature for industrial forecast systems. In the airline industry, a reliable forecasting of a demand for tickets at different fare levels forms a crucial step in a global optimization process, the objective of which is to sell a restricted number of available seats in a plane with a maximized revenue. Due to continuously changing demand caused by seasonality, special events like holidays or fairs, changes in the flight schedules or changes of the political or cultural situation of a country, there is a need for robust, adaptive forecasting techniques able to cope with such changes. In this paper an overview of various adaptive mechanisms used in the new forecasting system of the Lufthansa Airline is presented.