Joint Air Traffic Flow and Capacity Management: Large Scaling Dataset

This dataset provides the industry-sized problem instances and experimental results for the publication "ASPaeroFlow: Decomposition Heuristics for Joint Air Traffic Flow & Capacity Management" (accepted at ATMOS26) by Beiser, Hecher, Musliu, Trausmuth, and Woltran. The problem instances were generated using the ASPaeroFlow-DataGenerator and were utilized to benchmark the ASPaeroFlow instance-space decomposition heuristic, which aims to solve large-scale joint Air Traffic Flow and Capacity Management (ATFCM) problems by combining dynamic airspace configuration and traffic flow management. Dataset Contents 01_INSTANCES_SUPPLEMENTARY: Parsed, unparsed, and configuration files for the large instances. 02_OUTPUT_SUPPLEMENTARY: Experimental results for all benchmarks of the ASPaeroFlow paper. Instance Structure The benchmark set in 00_instance.zip is categorized by geographic regions of increasing size (e.g., CE, USA, EUR, DACH), encompassing graphs up to tens of thousands of vertices (navpoints) and exceeding 31,000 flights per instance. The dataset also includes instances generated for nominal capacity scaling (from 100% down to 10% capacity) to stress-test the algorithms. Each individual parsed instance directory contains the following parsed CSV files defining the ATFCM problem: airplanes.csv: Airplane specifications (e.g., speed in knots). airports.csv: Airport vertex definitions. sectors.csv: Airspace sector definitions and capacities. graph_edges.csv: Navigation graph edges with distances (in meters). flights.csv: Flight trajectory data (position and discretized time). airplane_flight_assignment.csv: Relational mapping of flights to specific airplanes. navaid_sector_assignment.csv: Relational mapping of navigation aids to airspace sectors. Operational Transformation Each instance directory also includes a transform_manifest.json and a mappings directory. These components contain the structural metadata required to transform the parsed instance data back into an unparsed mapping for operational usage. Furthermore, the unparsed instances are directly available. Citation Request If you utilize this dataset in your research, please cite the associated ATMOS26 paper.