Particle Swarm Optimization Assisted by Surrogates

This repository proposes new approaches for global, nonlinear and gradient-free optimization that combine the advantages of particle swarm optimization (PSO) and Bayesian optimization. Baseline and inspiration for this work is credited to the article Directed particle swarm optimization with Gaussian-process-based function forecasting.

Read our introduction!

Installing PSOAS

• Download or clone the repository (git clone git@github.com:upb-lea/PSOAS.git)
• Go to the PSOAS directory and fetch the requirements: python -m pip install -r requirements.txt (numpy and Cython have to be preinstalled for the smt installation)
• Afterwards run: python -m pip install -e .

Installing CEC-2013 to use it in the evaluation framework

(only needed for testing with the CEC-2013 benchmark)

• Download from https://github.com/yyamnk/cec2013single or git clone git@github.com:yyamnk/cec2013single.git
• Go to cec2013single/cec2013single/cec2013_func.c line 91
• Insert the absolute path to cec2013_data (e.g.: PATH-TO-DIR/cec2013single/cec2013single/cec2013_data)
• After inserting the path make sure to recompile: gcc cec2013_func.c
• Go back to cec2013single and run: python setup.py build_ext --inplace
• CAUTION: Make sure that you adjusted the import in the example notebook to reflect your folder structure

Citing

Please use the following BibTeX entry for citing us:

@online{MVSW2021,
  author = {Marvin Meyer and Hendrik Vater and Maximilian Schenke and Oliver Wallscheid},
  note   = {Paderborn University},
  title  = {Particle Swarm Optimization Assisted by Surrogates},
  year   = {2021},
  url    = {https://github.com/upb-lea/PSOAS},
}