||Dynamic Airline Scheduling : Analyses, Enchancements and Experiments
||Hansen, Troels Gotsæd
||Larsen, Allan (Logistics Optimisation Group, Centre for Traffic and Transport, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
||Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
||It is a challenge to operate in today’s increasingly competitive airline market. A problem many airlines are facing is that of fluctuating passenger demand and that the
schedule and assigned aircraft capacity do not always coincide with the demand. To approach this challenge, we present a Dynamic Airline Scheduling (DAS) model developed by Jiang (2006), which changes the departure times slightly (called retiming) and reassigns the aircraft types (called refleeting) to adapt the schedule and capacity to better demand forecasts.
We extend the work of Jiang (2006) in three main ways. We have more realistic data available, we perform many enhancements to the modeling of the problem, and we
carry out new experiments. The data we have available is actual forecast, schedule, cost and crew data from Scandinavian Airlines. The enhancements to the modeling we make are, among others, improvements of the cost function, more realistic aircraft turning times, and better modeling of passengers preferred departure times. Our main experiment is analyzing when to perform DAS, where we use 9 different optimization points and test for 7 days in a week. We also present four special scenarios and we analyze the impact on crew pairings when performing DAS close to departure.
The results show a significant increase in profit when using DAS and that the largest potential is close to departure. Also, we show that crew must be taken into
consideration if performing DAS at this stage.
Creation date: 2007-02-09
Update date: 2007-03-28