Air cargo demand prediction

Abstract

The air cargo transportation system is a large and complex service system, in which demand forecasting is a key element in the master planning process essential for analyzing existing cargo flight schedules and identifying future facility requirements of air cargo companies. We propose a multi producer/consumer solution for predicting cargo demand of a specific airline in a given route and cargo load factor for its flight schedule. This solution considers each airline as a producer and the users of air cargo services as consumers, with each producer having no explicit communication with other producers /airlines. The solution can assist airlines to maximize the usage of available cargo capacity. A major airline often has 100 million pounds of weekly cargo lift capacity. With this volume of cargo, even the slightest improvement in the forecasting technique and cargo load factor is liable to have a major impact in overall savings, performance, and efficiency. Our model uses the weighted majority learning algorithm [1] with various predictors for predicting the future demand. Based on the predicted demand, available cargo capacity, and by applying various strategies, new cargo capacity plan is suggested, thereby improving the cargo load factor as well as the financial bottom line.