Robust Resource Planning for Aircraft Ground Operations

Yagmur S. Gök, Daniel Guimarans, Peter J. Stuckey, Maurizio Tomasella, Cemalettin Ozturk

Abstract: Aircraft turnaround scheduling and airport ground services team/equipment planning directly concern both the airport operator and service providers. We first ensure airport-wide global optimality by solving a resource-constrained project scheduling problem (RCPSP) for minimal overall delays. We then support decentralized allocation of teams/vehicles to flights, independently by each service provider. Either a multiple traveling salesman problem with time-windows (mTSPTW), or a vehicle routing problem with time-windows (VRPTW) are solved for this purpose, by taking advantage of both constraint programming (CP) and mixed integer programming (MIP) solvers. We also exploit these models in a matheuristic approach based on large neighborhood search used to reach good solutions in reasonable time for real-world instances. Unlike the classical VRP objective of minimizing traveling time, we maximize the total slack time between team visits, and show that doing this fosters robustness of the generated plans. We assess the robustness of solutions through a discrete-event simulation model, and conclude by validating our approach with data provided by a major ground handling company for a day of operations at Barcelona El Prat Airport.