Comparing Centralized and Rolling Horizon Approaches for Optimal Aircraft Traffic Control in Terminal Areas

Abstract

This work addresses the real-time problem of managing takeoff and landing operations during traffic disturbances at a busy terminal control area (TCA). An important objective of traffic controllers is the minimization of delay propagation, which may reduce the aircraft travel time and the energy consumption. To improve the effectiveness of air traffic monitoring and control in a busy TCA, this paper presents an advanced optimization-based decision support system and compares centralized and rolling horizon approaches. The possible aircraft conflict detection and resolution actions were viewed as aircraft timing and routing decisions. The problem was modeled by an alternative graph formulation (i.e., a detailed model of air traffic flows in the TCA) and solved by scheduling and rerouting algorithms. The paper also proposes a new mixed-integer linear programming (MILP) formulation to compute (near) optimal scheduling and routing solutions. The paper compares the first-in, first-out (FIFO) rule, used as a surrogate for the dispatchers’ behavior, a truncated branch and bound algorithm for aircraft scheduling with fixed routes, a tabu search algorithm for combined aircraft scheduling and rerouting, and the MILP formulation solved via a commercial solver. Computational experiments are presented for practical-sized instances from Milano Malpensa Airport in Milan, Italy. Disturbed traffic situations were generated by simulating various sets of delayed landing and departing aircraft. A detailed analysis of the experimental results demonstrates that the solutions produced by the optimization algorithms are of a remarkably better quality compared with the FIFO rule, in relation to delay and travel time minimization.