The purpose of my research was to investigate if the cleaning process of Asito Transport Aircraft Services B.V. could be improved, and this was done with the following main question: 

Can the cleaning process of Asito Transport Aircraft Services B.V. be improved and what would the consequences of these improvements be?

Asito is a large cleaning company in the Netherlands that is spread out over the entire country and has several specializations, including Asito Medical and Asito Transport. Asito Transport Aircraft Services B.V. is a part of Asito Transport and is established at Schiphol Airport, which was also the location of this research study.

Asito Transport Aircraft Services B.V. is responsible for the cleaning of aircrafts of multiple airline companies. This process is bound by multiple restrictions; the most important being that the aircraft have to be cleaned on time and that Asito Transport Aircraft Services B.V. wishes to keep the total costs as low as possible.

Following an extensive analysis of the current situation it became clear that a scheduling method needed to be found which makes it possible to put the available capacity of the cleaning personnel to better use as well as have a better approach to the several external delay factors which influence the cleaning process of Asito Transport Aircraft Services B.V.. These external delay factors usually occur unexpectedly and are thus very hard to take into account when making the cleaning schedule.

A more flexible approach when dealing with the delay factors will also indirectly lead to a more efficient use of the available capacity, as the waiting time of the cleaning crews will become smaller. Improving these two aspects will most probably lead to a need for less personnel, which in turn will lead to a reduction of the total costs.

For Asito Transport Aircraft Services B.V. the best available scheduling method is a combination of two other methods: Multi-Server Queueing and Processor Sharing. The combination of these two methods results in the Multi Server Processor Sharing (MS-PS) theory: This is a scheduling method in which

  • servers (a cleaning crew split up into two sub-crews consisting of 5 people handle
  • multiple tasks (the aircraft which Asito Transport Aircraft Services B.V. has to clean)
at the same time to put the available capacity to the most efficient use.

The use of a simulation model made it is possible to research the influences of applying the Multi Server Processor Sharing method in the cleaning process of Asito Transport Aircraft Services B.V.. The following conclusions were made concerning the optimal use of the MS-PS scheduling method:

  • As queuing strategy for the waiting aircraft the use of ground time as the priority gave the best results.
  • Most of the cleaning personnel of Asito Transport Aircraft Services B.V. have to be enlisted during the morning hours as these are the peak-hours of the cleaning process.
  • The best results are achieved when all the available cleaning crews are put to work using the MS-PS method.
  • The optimal number of sub-crews is 2, where each of the sub-crews consists of 5 people.

A rendition of the simulation model used in this research project:

simulationmodel_cleaningprocess

The results of this research project show that the MS-PS method shows improvements for the cleaning process of Asito Transport Aircraft Services B.V. when compared to the current situation on the following points:

  • Improved use of the available capacity

By decreasing the number of people in a cleaning crew the available capacity is better put to use.

  • Equal handling of the aircraft

When more cleaning crews are available the average waiting time for the start of the cleaning process of the aircraft is greatly reduced as a result of the use of the MS-PS method. This means that aircraft get a more equal treatment.

  • Reduction of the influence of external delay factors

By using the MS-PS method the total time that cleaning crews are waiting is reduced with 60%. This means that the external delay factors have less influence in the total process time of the cleaning process.


Eveline Peijster MSc in Operations Research University of Amsterdam.