Increasing Flexibility of Airline Crew Schedules

AbstractAirline transportation frequently has to deal with disruptions, like technical breakdowns or bad weather conditions. Resulting delays often lead to additional delays, due to interdependencies between different resources such as aircraft or crews. In the following schedules may become infeasible during the day of operations. This forces the operations control to recover those schedules by mostly expensive actions. Regarding this difficulty robust scheduling deals with the construction of schedules that are less affected by disruptions, e.g. by increasing delay tolerance or providing possibilities for low-cost recovery actions. Robustness consists of two aspects, stability and flexibility. Stability describes the degree of the ability of a schedule to remain feasible and cost efficient under different operational environments without major modifications. In contrast, flexibility is the ability of a schedule to be adapted to changing environments in operations at low costs. However, increasing robustness accompanies with an increase of the planning costs. Thus a main goal of robust scheduling is its efficiency. In this context a method is efficient, if it obtains a high increase of robustness for a low increase of costs. In this paper we present a stochastic optimization model with respect to swapping opportunities of crews. We propose a run-time saving solution approach to integrate flexibility indicators into a column generation framework with a constraint generation extension. In order to realistically evaluate the robustness of the schedules we use a scenario based simulation model with enabled recovery actions for a realistic evaluation. The results show that the approach allows us to generate crew schedules with higher flexibility degrees at low increase of the planned costs in comparison with cost-efficient schedules