This repository includes the source code, the data and the plots for the paper entitled "A Scalable Divide-and-Conquer Genetic Algorithm for Multi-Objective Service Placement in the Cloud-Edge-Fog Computing Continuum", submitted for peer review to the journal "XXXX":

Carlos Guerrero, Isaac Lera, A Scalable Divide-and-Conquer Genetic Algorithm for Multi-Objective Service Placement in the Cloud-Edge-Fog Computing Continuum, XXXX, Volume X, 2026, Pages X-X, ISSN XXXX, https://doi.org/10.xxxx/xxxx.

If you would consider to cite the paper, you could use this bibtex record:

@article{GUERRERO2026,
title = {A Scalable Divide-and-Conquer Genetic Algorithm for Multi-Objective Service Placement in the Cloud-Edge-Fog Computing Continuum},
journal = {XXXX},
volume = {X},
pages = {X-X},
year = {2026},
issn = {XXXX},
doi = {https://doi.org/10.xxxx/xxxx},
url = {https://www.sciencedirect.com/science/article/pii/XXXX},
author = {Carlos Guerrero and Isaac Lera},
keywords = {Compute continuum, Genetic Algorithm, Problem Decomposition, Resource optimization, Multi-objective optimization, Application placement}
}

Directory Structure:

• src/: Contains all Python scripts for the simulation algorithm, NSGA-II execution, hierarchical scheduler, and plotting generators.
• config/: Contains the JSON configuration files representing the three experimental workloads (simulation_config_A.json, simulation_config_B.json, simulation_config_C.json).
• results/: Default output directory where each new simulation run generates its respective logs and metrics subdirectories.
• open_results/: Stores consolidated metric results (e.g., open_results/article_data/) specifically required for regenerating the plots and tables featured in the paper.

Reproducibility:

Three different experiments (A, B, and C) are studied in the article.

To reproduce the three experiments, execute the main.py script from the src directory, passing the respective configuration JSON file as an argument:

cd src
python main.py ../config/simulation_config_A.json
python main.py ../config/simulation_config_B.json
python main.py ../config/simulation_config_C.json

Generating the Plots:

To generate the JCR-ready PDFs corresponding to the paper's figures (multi-objective metric boxplots and Pareto front horizontal comparisons), simply execute the plotting scripts from the src directory.

They will automatically read the metrics stored in ../open_results/article_data/ and generate the PDFs in that same folder:

cd src
python generate_pdf_for_bloxplot.py
python generate_pdf_for_visual_comparison_paretos.py