Stochastic Ground-Delay-Program Planning in a Metroplex

This paper introduces a novel stochastic approach for the ground-delay-program planning under uncertainty, using chance-constrained optimization. The major advantage of the chance-constrained model is the ability to provide robust solutions with user-defined service level. The approach is compared with the Ball et al. (“A Stochastic Integer Program with Dual Network Structure and Its Application to the Ground-Holding Problem,” Operations Research, Vol. 51, No. 1, Feb. 2003, pp. 167–171) model for selecting planned airport acceptance rates for airports in a metroplex, which is an interdependent system in close geographic proximity. The approaches were evaluated using real flight schedules and landing-capacity data from the New York City metroplex airports. Although the Ball et al. model was found to be more efficient, the chance-constrained model shows the ability to provide a quantized way to balance the solution’s robustness and potential cost by choosing a proper service level. Moreover, the parallel-computing framework was demonstrated to be helpful in improving the computing efficiency, which suggests that deploying more computing resources would help solve a large-scale planning problem under uncertainty in the same framework.