A column-generation-based heuristic for passenger recovery problem considering multi-flights and seat-sharing constraints

Abstract

When facing flight delay or cancelation, airlines need to reassign affected passengers to other flights. In this paper, we model the passenger recovery problem as an integer program considering multi-flights and seat-sharing constraints. Correspondingly, we present a branch-and-price heuristic based on column generation framework to solve the passenger recovery problem. The framework consists of a master problem for selecting recovery schemes for passengers and subproblems for generating recovery schemes. Multi-flights constraints are considered in the subproblem, and seat-sharing constraints are considered in the master problem. Experimental results demonstrate that our method solves the passenger recovery problem with quite small gaps and significantly improves the satisfaction of passengers while balancing the financial cost of airlines within reasonable computing time.