The Assignment-Based-Anticlustering Algorithm - Computational Results

The Assignment-Based-Anticlustering (ABA) Algorithm is a scalable algorithm for anticlustering in Euclidean spaces. This repository provides detailed results from the computational study that was conducted in the following paper:

Baumann, P., Goldschmidt O., Hochbaum D.S., Yang J. (2026). A Fast and Effective Method for Euclidean Anticlustering: The Assignment-Based-Anticlustering Algorithm. https://arxiv.org/abs/2601.06351.

Detailed Results: Comparison to Leading Algorithms for Euclidean Anticlustering

The detailed results from the comparison to leading algorithms for Euclidean anticlustering are avaiable below:

Link to tables

Addtitional Results: Application of ABA to Euclidean Anticlustering with Categories

The detailed results from the comparison to leading algorithms on anticlustering with categories are avaiable below:

Link to results

Addtitional Results: Application of ABA to Balanced K-Cut

The detailed results from the comparison to METIS are avaiable below:

Link to results

Source Code of the ABA Algorithm

The Python source code of the ABA algorithm can be found here. Please note that the experiments in the paper were conducted using a C++ implementation of the ABA algorithm. Please contact Philipp Baumann (philipp.baumann@unibe.ch) if you are interested in the C++ version.

Reference

Please cite the following paper if you use the ABA algorithm.

Bibtex:

@misc{baumann2026fasteffectivemethodeuclidean,
      title={A Fast and Effective Method for Euclidean Anticlustering: The Assignment-Based-Anticlustering Algorithm}, 
      author={Philipp Baumann and Olivier Goldschmidt and Dorit S. Hochbaum and Jason Yang},
      year={2026},
      eprint={2601.06351},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2601.06351}, 
}

License

This project is licensed under the MIT License - see the LICENSE file for details