Team-oriented Airline Crew Rostering for Cockpit Personnel

Airline crew scheduling is a comparably well-studied field in Operations Research. An increasing demand for higher crew satisfaction arises; especially after most relevant cost factors have been optimized to their greatest extent, mostly with secondary or little regard on quality-of-life criteria for the involved crew members. One such criterion is team-orientation. Independent from the chosen assignment strategy (bidline systems, personalized rostering or preferential bidding) current approaches do not consider frequently occurring changes within daily or day-byday team compositions. By this, crew members rarely know with whom they work for the next flight(s) and/or day(s), respectively. In case of overnight stays outside their individual home base, crew members easily experience themselves having to find their ways to the booked hotels on their own. The avoidance of both aspects is highly appreciated by the crew as well as by the airlines, and will be addressed in the Team-oriented Rostering Problem (ToRP). In this work we present a first interpretation of Team-oriented Rostering (ToR) for cockpit crew, namely captains and first officers which can be implemented via two dedicated optimization models: Extended Rostering Model and Roster Combination Model. Due to the high combinatorial complexity certain strategies are applied during roster generation and roster combination in order to solve midsized instances based on a European tourist airline setting. As a result, the implied trade-off curve between operational cost and the number of team changes will be discussed.