https://github.com/bbopt/dmultimadspb
The DMultiMadsPB algorithm by Ludovic Salomon for multiobjective constrained blackbox optimization
Science Score: 13.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
○.zenodo.json file
-
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (10.9%) to scientific vocabulary
Repository
The DMultiMadsPB algorithm by Ludovic Salomon for multiobjective constrained blackbox optimization
Basic Info
- Host: GitHub
- Owner: bbopt
- License: lgpl-2.1
- Language: C++
- Default Branch: main
- Size: 248 KB
Statistics
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
DMulti-MADS (PB, TEB and Penalty variants)
This repertory contains the source code of the DMulti-MADS algorithm for constrained blackbox optimization. In terms of performance, it is more efficient than the old implementation found in [DMulti-MADS][https://github.com/bbopt/DMultiMadsEB].
Jean Bigeon, Sébastien Le Digabel and Ludovic Salomon, Handling of constraints in multiobjective blackbox optimization
Warning : This code has no vocation to be used in industry, see Nomad for a more robust implementation of state-of-the-art blackbox method. It aims at guaranteeing the reproducibility of the experiments described in this work.
Use
To use DMulti-MADS, Julia >= 1.6 is required. One can test it by typing the following command at the root of the directory. ```` julia> ]
(@v1.6) pkg> activate .
(DMultiMadsPB) pkg> test ```` All tests should pass.
A simple example can be found in the examples/ folder. One can look also at ./test/madsmodel.jl for more examples.
Problems
This folder contains an implementation of all multiobjective benchmark problems used in this article for the Nomad (BiMADS) software.
The algorithm [DFMO][http://www.iasi.cnr.it/~liuzzi/DFL/] is provided with all benchmarks coded in Fortran by the authors. For this reason, it is not given here.
Warning The generation of analytical benchmarks takes a lot of time (around three days and requires more than 40 G of memory on hardware;
For BiMADS, all executables are given with models and nelder-mead search deactivated. Uncomment the lines in the main function if you need them.
To obtain the real blackbox optimization applications, one can get them at: - [STYRENE][https://github.com/bbopt/styrene] - [SOLAR][https://github.com/bbopt/solar] You can compile them and launch them on the different algorithms with the scripts provided (minus paths to adapt to your machines):
Warning Solving STYRENE and SOLAR for a given solver takes one day.
Owner
- Name: Blackbox Optimization
- Login: bbopt
- Kind: organization
- Email: nomad@gerad.ca
- Location: Montréal, Qc, Canada
- Website: https://www.gerad.ca/nomad
- Repositories: 9
- Profile: https://github.com/bbopt
GitHub organization of the research group from Polytechnique Montréal and GERAD for derivative-free and blackbox optimization