The Lufthansa flight to Miami leaving from Frankfurt Airport’s terminal 1 – area A/B has been delayed. The flight has not received permission to get rolled onto the runway by the push back vehicles because a plane leaving for Rome just left its gate 10 minutes late and another plane arriving from Berlin needs to be brought into its proper position first. This scenario unfortunately happens remarkably often at these terminal areas for various reasons. These delays, known as Push-Back-Delays, have an immense influence on the punctuality of flights at Deutsche Lufthansa AG the head carrier at the airport of Frankfurt.
The analysis of these dependencies was the subject of Patrick Hünleins’ final paper. In cooperation with the Department of Materials Handling and Logistics Systems of the Karlsruhe Institute of Technology and the punctuality management team of the Frankfurt Division of Lufthansa AG, Patrick Hünlein determined the causes of these frequent Push-Back delays at the terminal area A/B. Based on his conclusions it was possible to create a new way to reduce the delays.
On average, 350 Lufthansa operated flights depart daily from Frankfurt Airport. Approximately 20-25% of these departures leave from terminal area A/B. In 2008, 3.2 % of the total delayed minutes attributable to Lufthansa AG were caused by Push-Back-Delays in this area. At first this percentage seems small but if you consider that a mere 1 % reduction in delayed minutes causes a huge improvement in air traffic efficiency and prevents cost overruns it is worth studying the underlying causes for this kind of delay.
Put simply, a Push-Back-Delay happens as soon as an airplane is ready for departure but cannot leave its parking position because another plane is operating in the area needed. The accurate determination of the reasons that cause this blockade situation first requires an analysis of the “real system airport”.
Hünlein determined that planning, operational procedure and infrastructure all have a direct influence on the Push-Back delays. In regards to planning, difficulties often result from specifications about the positioning of specific flights at specific gates and also from restrictions which limit the optional allocation of the flights to the gates. Infrastructural shortcomings contribute to Push-Back-Delays since the inner courtyard is a blind alley as you can see in picture No 1. In total there are only 14 gates available to accommodate the airplanes. Last but not least there are operational reasons that cause the delays such as interruptions in the dispatch process as well as the high number of plane movements in the terminal area.
There is never only one reason that causes the Push-Back-Delays but rather it is the combination of temporal, local and organisational value of every reason. The delays therefore depend on several causes. Further analysis isolated different factors which could, through their direct influence, reduce the Push-Back-Delays. These factors are:
1. The premises of planning (planning)
2. The volume of traffic (operational)
3. The infrastructure of the terminal area itself (infrastructural)
The next step was putting these discoveries in a model. Therefore Patrick Hünlein used the Airport Gate Assignment Problems (AGAP). These optimized problems are a special case of square allocating models in which, keeping several criteria in mind, the scientist attempts to allocate a given number of flights to a given number of gates in the most efficient manner, often there are fewer gates than flights. In this model the developed dependencies were considered to identify the real reasons using different scenarios. Furthermore, it was possible to calculate acceptable schedules with a minimum of Push-Back-Delays as well as to determine the reasons that influence the delays. These schedules are supposed to be the decision basis for positioning the flights in the future. Beyond that it should be possible to identify the detailed reasons for the Push-Back-Delays.
In the target function there were two different factors considered. On one hand the aim was to minimize the Push-Back-Delays which are caused by the positioning and on the other hand to maximize the compliance with the premises of planning as far as possible. The number of Push-Back-Delays was identified with an illustration of blockade situations (figure No.2). Using the premises of planning it was necessary to revert to a “punishment function”. This function sets additional minutes of delay in the target function of flights which haven’t been positioned in the alley. Furthermore, in the constraints all restrictions were taken into consideration. Some examples are the requirement that two planes cannot simultaneously occupy the same gate nor can smaller gates accommodate big planes like the Airbus 380 or the Boeing 747.
For the determination of authorized occupancy plans this optimization problem was transformed into a suitable software environment and was solved by integrated solution algorithms. Several scenarios were investigated to determine how the influence of the previously found values affected the number of the Push-Back-Delays which occurred. At the point of traffic volume for example the rate of occupancy, and thus derivationally the amount of aircraft movements, was limited to within a range of 20-30%. Within the tests of the premises of planning the set punishment function was used. In one scenario testing premises of planning high “punishment points” were set e.g. when an aircraft was not placed in the inner courtyard. This caused the target function get worse.
It became evident that the premises of planning – the guidelines setting at which gate which aircraft is operated – are the primary cause of Push-Back delays. The amount of flight movements amplifies this influence, causing delays more frequently at higher volumes. A further step clarified if this knowledge could be applied beneficially to the real system. In a simulation case study the found results could be confirmed. Thereby it was determined how the calculated occupancy plans apply in reality and how the Push-Back delays change under these conditions. So, delays and advancements were assigned to the airplanes in order to adjust them more to the real times of departures and arrivals. The pattern of these adjustments matched the actual times which are prevailing in flight traffic. In this way it was taken into consideration that the flight traffic rarely conforms to the scheduled timetable. In this manner, the previously determined dependencies could be validated and can be safely said to be confirmed.
Based on this knowledge a procedure could be worked out how the plans of occupancy can be improved in the future in order to minimize Push-Back delays. With the help of the AGAP model the other potential factors which can contribute to an immediate improvement can be shown.
1. More flexibility in positioning the airplanes.
2. The amount of traffic should not exceed a defined level within a longer period of time than currently ordained
3. Flexibility to rapidly adjust occupancy plans if the flight schedule changes.
The first two premises of action have the biggest potential to reduce Push-Back delays because they are directly in relation to the causes of the delay. It can be guaranteed that despite the dynamic aspects of airport delays there always exist optimal plans of occupancy which minimize Push-Back blockades.
With the analysis of Push-Back dependencies a partial contribution to the reduction of delays in the total system airport was worked out. The thing to determine therefore is which possible effects on the directly affected areas could result by the presented procedure. If – on the one hand – improvements are achieved but – on the other hand – degradations result, the exact effects have to be known already before the application of the new procedures. Thus, the next step would be a further development of this topic analyzing which additional key factors influence punctuality of airline traffic besides Push-Back delays.
Within the scope of this work a new model for the determination of gate occupancy plans with minimal Push-Back-Delays was developed and verified by a simulation study. Hence, the primary goal of this work was achieved: The identifications of the Push-Back dependencies and their results as well as the implementation of a suitable procedure which makes a contribution towards the reduction of Push-Back delays.