Solving a large-scale integrated fleet assignment and crew pairing problem

Abstract

Airline schedule planning problems are typically decomposed into smaller problems, which are solved in a sequential manner, due to the complexity of the overall problems. This results in suboptimal solutions as well as feasibility issues in the consecutive phases. In this study, we address the integrated fleet assignment and crew pairing problem (IFACPP) of a European Airline. The specific network and cost structures allow us to develop novel approaches to this integrated problem. We propose an optimization-driven algorithm that can efficiently handle large scale instances of the IFACPP. We perform a computational study on real-world monthly flight schedules to test the performance of our solution method. Based on the results on instances with up to 27,500 flight legs, we show that our algorithm provides solutions with significant cost savings over the sequential approach.