/SLS_Evidence_long_tail

Supplementary data of the paper "Evidence for Long-Tails in SLS Algorithms" (ESA 2021) by Florian Wörz and Jan-Hendrik Lorenz.

Primary LanguageJupyter NotebookOtherNOASSERTION

Supplementary data of "Evidence for Long-Tails in SLS Algorithms"

This repository contains all data produced for the empirical evaluations of the paper "Evidence for Long-Tails in SLS Algorithms" of Florian Wörz and Jan-Hendrik Lorenz.

The instances used to obtain this data can be found in the Zenodo repository specified in the paper: https://doi.org/10.5281/zenodo.4715893.

General Folder Structure

The repository contains the following directories/files. Here, we briefly describe the contents of each folder. Furthermore, all important folders contain a description file which gives more details.

  • ./evaluation

    Contains the evaluation (via Jupyter Notebooks) of the collected data. This is the most important folder. All evaluations take place in the files ./evaluation/jupyter/evaluate_*.ipynb.

  • ./experiments

    Contains all csv data files used for analysis. These files constitute the aggregated raw data.

  • ./scripts.tar.xz

    Contains scripts used for data clean-up of the original collected data and for generating the instances.

Authors

Florian Wörz, Universität Ulm, Institut für Theoretische Informatik, Ulm, Germany

Jan-Hendrik Lorenz, Universität Ulm, Institut für Theoretische Informatik, Ulm, Germany

Acknowledgments

The authors acknowledge support by the state of Baden-Württemberg through bwHPC.

Florian Wörz was supported by the Deutsche Forschungsgemeinschaft (DFG) under project number 430150230, "Complexity measures for solving propositional formulas".

License

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