Simulation-based key performance indicators for evaluating the quality of airline demand forecasting

This article describes an approach to evaluating the quality of airline demand forecasting. It presents a a simulation framework that includes a detailed model for generating artificial demand. In this system forecasting methods can be compared in a stable, controllable environment. Their performance may be rated based on the overall system output in terms of revenue and bookings as well as through common error measurements. In addition, the use of a psychic forecast as a benchmark is proposed and illustrated by first results.