Airspace Sectorization by Dynamic Density

We propose and analyze an algorithm for airspace sectorization that uses Dynamic Density (DD) as the objective function. A primary hypothesis of this work is that the explicit use of Dynamic Density as the objective function is required to generate airspace sectors that consider all appropriate aspects of controller workload and cognitive complexity in managing the traffic in the sector. Due to the nature of the DD metric, which relies on explicit numerical computation of factors related to airspace sector boundary locations, traffic trajectories, and the inter-relationships between traffic and boundaries, it is necessary to use an optimization method that is compatible with general numeric objective functions. Although evolutionary computing might provide a greater likelihood of near-optimal results, the processing requirements of computation of the full DD metric in a realistic problem environment are prohibitive. A heuristic gradient descent approach has been implemented. The analysis results demonstrate the efficacy of the approach in the generation of airspace partitions and the successful manipulation of DD results in the sectors that are created by the algorithm.