Increasing Stability of Crew Schedules in Airlines

In airline traffic disruptions occur frequently and cannot be totally avoided. They may lead to infeasible aircraft and crew schedules during the day of operations, due to absence of resources or violation of crew rules. The process of finding new schedules in such cases is called recovery or disruption management. The short-term recovery actions usually imply additional costs meaning that the total operational costs of a crew schedule can be significantly higher than the original planned costs. It is generally desirable to construct the schedule already in the planning phase in such a way that not just the planned costs, but the total operational costs are minimized. The goal is thus to construct schedules which remain feasible or can be recovered without high costs in cases of disturbances. This approach is generally called robust scheduling. There is no unique definition about robustness in airline scheduling. From our point of view it is important to realize that robustness involves the two aspects stability and flexibility. A schedule is stable if changes in operating environment imply only small changes in the schedule. For example delays can to a large extent be absorbed by buffer time at the right place without propagating them further. On the other hand, a flexible schedule involves enough possibilities for schedule changes so that deviations in the operating environment can be recovered efficiently. For example it is possible to swap crew or aircraft, or use a reserve crew to recover from a delay. Stability can often be increased with a low cost, for example by replacing buffer selectively in places that are at risk of being subject to disruptions. A higher stability or flexibility can be achieved by increased cost for example with additional crews and aircraft. In this study we focus on actions to increase stability with low cost since these are generally desirable and worth of realizing. The consideration of disruption management or robustness is subject of publications since late 1990s when a fast growth of air traffic caused increased congestion. Mercier et al. (2005) propose an integrated aircraft routing and crew scheduling problem incorporating a robustness measure. For the robustness measure the available ground time for crews during an aircraft change is considered. Weide et al. (2007) propose to solve the integrated aircraft routing and crew scheduling model heuristically by solving aircraft routing and