Design and simulation of airport congestion control algorithms

This paper proposes a stochastic model of runway departures and a dynamic programming algorithm for their control at congested airports. Using a multi-variable state description that includes the capacity forecast, the runway system is modeled as a semi-Markov process. The paper then introduces a queuing system for modeling the controlled departure process that enables the efficient calculation of optimal pushback policies using decomposition techniques. The developed algorithm is simulated at Philadelphia International Airport, and compared to other potential control strategies including a threshold-policy. The algorithm is also shown to effectively adapt to changes in airport departure capacity, maintain runway utilization and efficiently manage congestion.