Airlines need crew to operate their flights. In case of crew unavailability, for example due to illness, the airline often uses reserve crew to still be able to operate the flight. In this paper, we apply a simulation-based optimization method to determine how much and on which days reserve crew needs to be scheduled. This method is a combination of reserve optimization and reserve evaluation. The optimization model creates sets of reserve crew that are checked by the evaluation model using simulations of disruptions of the flight schedule. On top of this, we introduce the concept of mixed reserves as a new type of reserve crew scheduling with variable duties and length. Using flight, disruption and reserves data from a major European airline, it is found that using our method leads to fewer reserve crews. Based on the simulation of a large set of schedule disruptions, it was observed that we would need 12% less working days allocated to schedule recovery.
In the proceedings of the 23rd ATRS World Conference, Amsterdam (The Netherlands), 2-5 July 2019