Constructing large-scale benchmark datasets for the hierarchical hub location problem using geospatial information: a...

Area of Science:

• Operations Research
• Geographic Information Science
• Urban Planning

Background:

• Lack of large-scale, realistic benchmark datasets for the single-allocation hierarchical hub median problem hinders research in location science.
• Existing datasets are often not publicly available or lack the granularity needed for robust validation.

Purpose of the Study:

• To introduce an innovative methodology for generating large-scale, realistic benchmark geospatial datasets.
• To address the scarcity of public datasets for the single-allocation hierarchical hub median problem.
• To provide reproducible and scalable datasets for algorithmic benchmarking and urban planning.

Main Methods:

• Developed a step-by-step methodology for dataset generation.
• Utilized actual building-level geographic data from OpenStreetMap and QGIS.
• Created datasets based on Kolkata and Mumbai, India's major metropolitan areas.
• Solved benchmark instances using IBM ILOG CPLEX for exact solutions.

Main Results:

• Successfully generated diverse, large-scale benchmark geospatial datasets.
• Provided ready-to-use datasets for the single-allocation hierarchical hub median problem.
• Ensured datasets are reproducible, scalable, and adaptable to various problem formulations.
• Empirical validation of algorithms is supported through exact solutions.

Conclusions:

• The public availability of these datasets fills a critical research gap in location science.
• The datasets serve as a foundational platform for benchmarking algorithms and advancing hierarchical hub location studies.
• This work facilitates future methodological innovations in urban planning and logistics optimization.