Papers that use Brave New Algorithm for different purposes, mainly measuring energy expenses; it uses code at BraveNewAlgorithm.jl and is hosted at a GitHub repository.
The Brave New Algorithm is an evolutionary algorithm that constraints reproduction according to a system of castes inspired in the Brave New World novel by Aldous Huxley. That way it can control more effectively the exploration-exploitation balance. It is fully described in this paper
These are the sources of the papers accepted so far
Citation:
JJ Merelo, Cecilia Merelo Molina Best practices in measuring energy consumption in population-based metaheuristics, in Proceedings OLA’26 International Conference on Optimization and Learning, pp 183-194, available online: https://vb.svako.lt/object/elaba:291738676/291738676.pdf#page=188.
Check out the source code at ola-26.Rnw. This was created for the
OLA 26 conference in
Chania, Crete.
ola-26.R contains just the
code used to generate tables and graphs.
Check out the presentation.
Additional content has been created: ola-26-explainer.Rmd is a
summary of the main findings. Read it
online here, or render it locally with
rmarkdown::render("ola-26-explainer.Rmd") after cloning the
repository.
An explanation of how operational context influences the workload measurements is shown in this file, generated by this markdown file.
The paper is published in the OLA’26 proceedings, but there’s also an anotated version that can help you navigate through the main findings and how data is processed.
Content for this paper is included in this release.
Paper source: lion-26.Rnw
with extracted R code in lion-26.R.
A short summary of the main findings in the paper.
rmarkdown::render("lion-26-explainer.Rmd") if everything is
installed properly.An annotated version of the paper with the main findings and other highlights is included in the release.
A critical examination of the experimental data by modeling energy consumption in terms of the algorithm parameters.
A companion document traces the research arc from LION-26 to OLA-26: what the first paper found, which conclusions the second paper revised, and why (with live R plots built from both datasets):
rmarkdown::render("lion-ola-progression.Rmd") after cloning
the repository.Please reference one of these papers when re-using the content of this repository, code or data
@misc{bna25,
year = {2025},
month = {11},
url = {https://hdl.handle.net/10481/107864},
abstract = {Green computing tries to push a series of best practices that, in general, reduce the amount of energy consumed to perform a given piece of work. There are no fixed rules for {\em greening} an algorithm implementation, which means that we need to create a methodology that, after profiling the energy spent by an algorithm implementation, comes up with specific rules that will optimize the amount of energy spent. In population based algorithms, the exploration/exploitation balance is one of the most critical aspects. The algorithm we will be working with in this paper called Brave New Algorithm was designed with the main objective of keeping that balance in an optimal way through the stratification of the population. In this paper we will analyze how this balance affects the energy consumption of the algorithm.},
organization = {Ministerio español de Economía y Competitividad: proyecto PID2023-147409NB-C21.},
keywords = {Green Computing},
keywords = {Energy profiling},
keywords = {Metaheuristics},
title = {Analyzing how the exploration/exploitation trade off in biologically-inspired algorithms affects energy consumption},
author = {Merelo Guervos, Juan Julián and Merelo-Molina, Cecilia},
}This is open science. A bit of orientation
data
contains the results of the different experimentsLICENSE