Hyper-Heuristics Powered by Artificial Neural Networks for Cusomising Population-based Metahueristics in Continuous Optimisation Problems
This repository contains the resulting datasets of using the hyper-heuristic based on neural networks to produce enhanced metahueristics. We also included the main jupyter file to plot the figures thereby presented.
Authors: José Manuel Tapia-Avitia, Jorge M. Cruz-Duarte, Ivan Amaya, José Carlos Ortiz-Bayliss, Hugo Terashima Marín, Nelishia Pillay
Due to the file size limitation of GitHub, we provided the resulting data files in split zip files into data_files: all-exp-results.zip
, all-exp-results.z01
, all-exp-results.z02
, ..., all-exp-results.z33
. So, after cloning this repository, you must follow the steps below in your terminal.
- Go to the
tl-hh-umhs/data_files
folder:
cd data_files
- Combine the split zip files:
zip -F all-exp-results.zip --out all-exp-results-single.zip`
- Unzip the combined zip file:
unzip all-exp-results-single.zip
Then, you have the data result files.
- Python v3.8+
- CUSTOMHyS framework
- Standard modules: os, matplotlib, seaborn, numpy, pandas, scipy.stats, tensorflow
- Main notebook: processing_results_nnhh.ipynb
- Results from the proposed approaches (this folder will be available once unzip all-exp-results files): data_files/all-exp-results
- Raw figures generated from the main notebook: data_files/exp_figures
- Experimental configurations used in this work: exconf
- Results for basic metaheuristics: data_files/basic-metaheuristics-data_v2.json
- Collection of basic metaheuristics: collections/basicmetaheuristics.txt
- Collection of default heuristics: collections/default.txt
José Manuel Tapia-Avitia - [email protected] Jorge M. Cruz-Duarte - jcrvz.co, [email protected]
Distributed under the MIT license. See LICENSE for more information.