The air cargo industry provides fast, secure and reliable transportation of goods. A large variety of different goods from spare parts over cars to living animals such as horses are transported by air. In order to minimize the heterogeneity of air cargo shipments and save valuable time loading and unloading the aircraft, shipments that are booked on the same flight are consolidated onto unit load devices (ULDs). This process is called build-up process and is very space- and labor-intensive. The uneven distribution of workload as well as time and space restrictions make it very difficult to schedule the build-up processes and the right amount of manpower.
The objective of this thesis is to design forecasting methods in order to predict the amount of built-up relevant cargo and number of ULDs per vehicle category and day. An existing forecast that is reliable and very accurate but too coarse-grained is used as input parameter. The existing forecast is broken down in multiple steps using an out-of-sample approach. A discussion about the fitness of various forecast accuracy measurements is conducted. Evaluating the forecast model, we were able to retain almost 90% of the initial forecast accuracy.Die Teilnahme an den Vorträgen der KommilitonInnen ist Pflicht.