Premium
Operational Risk in Airline Crew Scheduling: Do Features of Flight Delays Matter? *
Author(s) -
Sun Xuting,
Chung SaiHo,
Ma HoiLam
Publication year - 2020
Publication title -
decision sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.238
H-Index - 108
eISSN - 1540-5915
pISSN - 0011-7315
DOI - 10.1111/deci.12426
Subject(s) - crew , computer science , operations research , aviation , scheduling (production processes) , punctuality , interdependence , robustness (evolution) , crew scheduling , operations management , aeronautics , transport engineering , engineering , biochemistry , chemistry , law , political science , gene , aerospace engineering
Our work is motivated by the increasing demand in the aviation sector and simultaneously aggravated poor punctuality. The airlines play an important role in improving their service level and mitigate the profit risks incurred due to their poor resources planning. However, to identify and mitigate operational risks faced by the airlines is complicated, as they are coming from both internal and external factors. Due to the realistic nature, we explore the flying time characteristics, and further model the consecutive interdependent departure‐arrival times. It is a key feature included in this study that has not been studied in literature. We characterize the flying time of each flight by the heteroscedastic regression model. The analytical closed‐form for the recursive relationship of the expected departure and arrival times of connective flight legs is then carried out. Accordingly, we propose a novel data‐driven bicriteria mathematical model in which the interdependent structures of the departure and arrival times of the consecutive flights is incorporated into the robust optimization. A column generation‐based algorithm is developed to solve the proposed model. We find that, for more than 23% of the flights explored, the expected flying times are significantly influenced by its actual departure times. The real‐data based computational examples identify that our proposed model sufficiently improves the reliability of the crew pairings decisions by reducing the total deviated time from the schedules with a slight increase of the total basic crew operations cost. Some managerial implications for robust crew pairing and determination of robustness level are discussed as well.