The crew split problem for team crew rostering

The Crew Pairing Problem (CPP) is an optimization problem at the heart of the aviation industry, focused on creating the minimum-cost set of crew pairings. These pairings are sequences of flights and rest periods that together form the work schedule for a crew over several days. The goal is to cover all flights within a given period while minimizing costs. Once the optimal pairings are identified, they are assigned to individual crews in a next step.

Airlines often benefit when cabin crews work together as a team for entire pairings. This approach not only boosts team morale and efficiency but also enhances the robustness of the flight schedule. However, varying crew requirements for different flights—such as the number of crew members and their specific roles on the plane—can make it challenging to keep a team intact across consecutive flights. In such cases, forming a core team to cover most of the requirements and supplementing it with additional crew members can be advantageous. Despite being recognized as a significant issue, finding the optimal way to form these teams has not been systematically addressed until now.

In this presentation, we introduce the Crew Split Problem (CSP), a novel approach to determine the composition of core teams for each flight and identify the complementary crews that will join them. The CSP is a multi-objective optimization problem that seeks to maximize the size of each core team, minimize connection times, and reduce deadhead costs. Additionally, our model incorporates the concept of downranking, allowing higher-ranked crew members to work in lower-ranked positions if necessary. We demonstrate the effectiveness of the Crew Split Problem through a large-scale real-world application, highlighting its potential benefits for the airline industry.