The main goal of my research was to provide more insight into how the resources in one of the freight buildings of KLM Cargo could be controlled to improve the efficiency of workload management.

KLM Cargo is the freight division of KLM, the Royal Dutch Airlines. The home base is at Amsterdam Airport Schiphol, where this research took place. It has a hub-function, where freight arrives and departs by truck as well as by aircraft. World Port, one of the three freight buildings of the KLM Cargo HUB, handles freight that has to be flown to an intercontinental destination. The research focused on this part of the KLM Cargo HUB.

Terminal manpower costs are the largest among all operating costs, so efficient management of manpower resources helps terminals to improve manpower utilization and reduce operating costs accordingly. Designing an efficient work schedule is a complex activity due to a lot of (stochastic) variables that need to be taken into account.

The management expressed the wish to have a tool for optimizing the manpower scheduling. The research pointed out this could be reached by developing a simulation model. Manpower scheduling is a complex activity and with limited information even very complex. The combination of the simulation model and the report, gives a support to the manpower scheduling.

In a freight terminal freight changes from one modality to another. For the transport of freight by aircraft special cargo carriers are needed. These standardized cargo units are called Unit Load Devices (ULDs) and they can be divided into three different types: T-ULDs, M-ULDs and BBs.

  • T-ULD: is filled with one destination and ready for flight.
  • M-ULD: this type should be broken down, due to the fact it contains shipments with multiple destinations.
  • BB ULD: not all the space on the ULD is used, there is some space left to add more shipments with the same destination.


At the freight building, freight can be delivered by truck at the Moving Truck Dock (MTD) entry (freight from Europe) or at the export entry (freight from the Netherlands) and by aircraft (freight from intercontinental destinations). After handling the freight it will be flown further to an intercontinental destination.

When we zoom further in into the processes at the freight building, there are beside uploading two other main processes recognizable; breaking down and building up of pallets. Depending on the type of freight it needs complete these processes or not. The time to perform these processes mainly depends on the number of employees working.

Using the simulation model results can be analyzed about:

  • the amount of packages or ULDs handled at each of the processes per hour/shift/day
  • the utilization at each of the processes per hour/shift/day
  • the staying time at each of the processes per hour/shift/day
  • the waiting time at each of the processes per hour/shift/day
  • the size of the queuing line at each of the processes per hour/shift/day
  • the overall on-time performance


Using all this information, the management can make well-considered decisions about the manpower planning. Important to notice is that the quality of the outcomes of the model depends totally on the quality of the input.


Four different scenarios were performed during this research, the main conclusions from these scenarios are:

  • If the number of employees grows, the performance (percentage freight that is delivered on time) goes up. This was a very likely conclusion, but more interesting was the result that the impact of less employees in the building is bigger than the impact of more employees.
  • If the amount of offered freight goes up, the performance goes down (assuming the number of employees is constant).
  • The T/M ratio represents the ratio between the amount of T-ULDs and the amount of M-ULDs delivered at the freight building. If the T/M ratio goes up, the performance will go up. This is caused by the fact that M-ULDs are the ULDs which causes most workload at the warehouse. Further, there could be concluded that less T-ULDs would have a real impact on the performance and more T-ULDs would affect just a small improvement on the performance.
  • At last, the time to start breaking down a pallet has been investigated. It appeared that the performance will increase slightly if the trigger to start breaking down a pallet will be longer before departure. The differences are most noticeable when the trigger is less than 5 hours before departure.


Linda Veldhuizen MSc in Operations Research University of Amsterdam.