An Optimisation and Forecasting Framework for ULD Packing in the Air Cargo Supply Chain

Abstract

The air cargo industry is a challenging environment due to the high competition between the stakeholders involved. This demands, as example, high efficiency from cargo airlines. Efficiency can be ensured by designing loading strategies that fully exploit the available cargo volume. Unknowns in the booking dimensions and flight information make this task challenging. We propose an optimisation and forecasting framework for ULD packing. Goal of the framework is to assist analysts with their decision-making, when assessing whether a new booking should be accepted or declined. Unknown booking dimensions and flight details are forecasted using an ensemble learning method. This information is used as input for a bin packing heuristic, which assesses if the booking fits in the aircraft. In addition, we propose a risk index that quantifies what are the chances the proposed loading configuration (per ULD) is not feasible due to a mismatch between predicted and real booking dimensions.