A Multi-Objective Generalized Random Adaptive Search Procedure for Resolving Airspace Congestion

This paper extends previous work that examines the utility of employing a Generalized Random Adaptive Search Procedure (GRASP) to minimize the impact on the National Airspace while resolving congestion. As today's methods for managing congestion are mostly manual, and for which predictions of both capacity and traffic demand are uncertain, it is often difficult to find efficient, flight-specific resolution maneuvers for the hundreds of flights affected when congestion arises. This work continues to investigate improvements upon previous research that develops a non-optimal flight-specific congestion solution strategy based on a prioritized flight list by iteratively perturbing this prioritization to find better solutions. In this paper we examine the previous prioritization criteria considered and develop a new prioritization criterion that incorporates the two individual criteria. The research shows that combining the two criteria effectively bridges the difference between the results produced by the single criteria resulting in consistently good solutions across multiple metrics and objectives. Following this, we extend the optimization framework to include multi-metric objectives, examining the trade-off of congestion versus delay and how including inequity in the objective affects all three metrics. The research shows that significant benefits can be derived even when a metric is only lightly weighted in the objective, showing that multi-metric objectives are a significantly desirable formulation.