Star Discrepancy Benchmark Results for Black-Box Optimization Algorithms

A repository containing code and data for reproducing experiments from the paper 'Computing Star Discrepancies with Numerical Black-Box Optimization Algorithms'. The data includes raw algorithm performance logs, processed convergence information, and final discrepancy values generated using the IOHexperimenter framework. The work was authored by François Clément of the Centre National de la Recherche Scientifique.

Use Cases

• Benchmarking optimization algorithms based on performance data on star discrepancy problems
• Analyzing algorithm convergence behavior based on the provided convergence information CSV files
• Reproducing computational experiments from the associated research paper using the provided code and raw data
• Visualizing algorithm performance and landscape plots using the included Python notebook

Strengths

• Includes raw data, processed results, and all code necessary for full reproducibility
• Provides final discrepancy values found by two specific algorithms (TA and DEM)
• Contains a relative performance comparison table for multiple optimizers

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Last update date is unknown; freshness unverified

Provenance

Source

Centre National de la Recherche Scientifique

Collection Method

Generated by running black-box optimization algorithms on discrepancy problems using the IOHexperimenter package.