This grant provides funding for the development of mathematical models that will serve to integrate the airline operational planning problems of fleet assignment, aircraft routing, and crew scheduling, and to design and test effective solution methodologies. Three models will be investigated for this purpose. The first will study the integration of flight scheduling and fleet assignment to simultaneously determine how many passengers to accept on each path using itinerary-based demands and considering airline recapture, flexible flight times, and schedule balance. The second model will adopt an alternative flight network modeling paradigm in lieu of the traditional time-space network to additionally integrate the design of aircraft routes, while taking heed of through-flights. The third model will consider a more comprehensive integration of the fleet assignment, aircraft routing, and crew scheduling processes. Exact and heuristic decomposition solution approaches will be developed, which will include polyhedral analyses for generating valid inequalities, stabilized branch-and-price-and-cut, Lagrangian relaxation, and model enhancements using the Reformulation-Linearization Technique in order to cope with the large-scale model formulations. The models and solutions procedures will be tested and validated through computational experiments on real-data provided by United Airlines and Tunis Air.
If successful, the results of this research will positively impact the profitability and service quality of airline companies. The principal goal of this project is to demonstrate the benefits of simultaneous consideration of schedule planning, aircraft fleet assignment and routing, and crew scheduling. By recognizing the interplay among these operational planning problems and integrating them within consolidated models, airlines will be able to make more profitable decisions over the traditional sequential process that examines these interrelated problems separately. The solution methodologies developed for addressing the large-scale models will also contribute toward the repertoire of concepts and approaches for solving other related scheduling and discrete optimization problems.
