Increasing stability of crew and aircraft schedules

Abstract

Stability describes the grade of the ability of a plan to remain feasible and cost efficient under variations of the operating environment without major modifications to the plan. In airline traffic delays often lead to additional delays, because of interdependencies between different resources such as ai rcraft, crews and airport facilities. In this paper we focus on reactionary delays that result from crews changing aircraft. Based on a stochastic model for delay propagation we propose an indicator for stability of airline crew and aircraft schedules. This indicator is used to generate more robust aircraft and crew schedules. An integrated formulation for the aircraft and crew scheduling problem results in a non-linear stochastic recourse function. We decompose this formulation into separate linear problems connected by the objective function. The decomposed stochastic problem may be solved using a heuristic iterative approach based on column generation and dynamic programming for the recourse functions. In order to evaluate the robustness of the generated schedules and to compare with a deterministic approach for robust scheduling we use an also presented simulation model.