Using Sector Complexity Metrics to Predict Sector Capacity

This paper discusses the prediction of sector capacity in the United States’ National Airspace System (NAS) using sector complexity features as input. Sector capacity prediction is critical for safe operation of aircraft in the NAS. Among other uses, a sector’s capacity is the basis of flow decisions designed to limit a sector’s traffic to that which can be safely separated. In this paper we use Machine Learning techniques with features from the Sector Complexity literature to predict capacity. We present the accuracy of the calculation, insights into input sensitivity, and recommendations for future studies.