This repository contains our work on Parameter Tuning and Genetic Improvement, which is submitted to the GI@ICSE2025 titled "Empirical Comparison of Runtime Improvement Approaches: Genetic Improvement, Parameter Tuning, and Their Combination".
This repository is structured as follows:
- MAGPIE-OUR-VERSION: contains our version of the MAGPIE
- _magpie_logs: raw outputs from the experiments.
- magpie: the main functionalities of the MAPIE including the algorithms.
- results: MiniSAT experiment results from 10 k-fold in csv format.
- appendix_algorithms_and_tables.pdf: contains the additional information, including are not limited to, pseudo-code of the algorithms, tables and graphs.
- Our MAGPIE version for this study: https://anonymous.4open.science/r/w6hlu50lds662C/
- Benchmark: http://aclib.net/cssc2014/solvers/minisat_HACK_999ED_CSSC.tar.gz
- Train + Test Instances: http://aclib.net/cssc2014/instances/circuit_fuzz.tar.gz
- MAGPIE base version for this study: https://github.com/bloa/magpie/tree/5b85d5dc11380abbfed5acaab33f8cc8f53d3f8e