Coefficient Mutation in GP-GOMEA

This repo contains the code for the paper:

@inproceedings{virgolin2022coefficient,
author = {Virgolin, Marco and Bosman, Peter A. N.},
title = {Coefficient Mutation in the Gene-Pool Optimal Mixing Evolutionary Algorithm for Symbolic Regression},
year = {2022},
publisher = {Association for Computing Machinery},
url = {https://doi.org/10.1145/3520304.3534036},
doi = {10.1145/3520304.3534036},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
pages = {2289–2297},
numpages = {9},
}

How to reproduce the experiments

You must install GP-GOMEA and SRBench to reproduce the experiments. See the respective READMEs. A copy of the GP benchmark collection by Luiz Otavio Vilas Boas Oliveira et al. is duplicated in this repo, no need to install this separately.

To reproduce the first part of the experiments, use run.py (runbatch.sh calls run.py). To reproduce the second part, you must use SRBench, see the README of SRBench's repo. To obtain plots and info, use analysis.ipynb.