A Comparison of Eight Optimization Methods Applied to a Wind Farm Layout Optimization Problem

By Jared J. Thomas, Nicholas F. Baker, Paul Malisani, Erik Quaeghebeur, Sebastian Sanchez Perez-Moreno, John Jasa, Christopher Bay, Federico Tilli, David Bieniek, Nick Robinson, Andrew P. J. Stanley, Wesley Holt, and Andrew Ning.

A manuscript submitted to Wind Energy Science.

This repository is for shared code, data, and other things, but not the manuscript (which is handled on overleaf).

• Optimization results can be found in the .yaml files in ./results/optimization-results/
• Conda environment used for this project can be found in conda-env.yml
  • create with $ conda env create -f conda-env.yml
  • activate with $ conda activate julia-env
  • this environment includes julia, so you don't need to install it directly.
• Image/table data and generation code can be found at ./image-gen/
  • The figure/table generation script for most of the figures and one table is in Julia.
    • Start a Julia session: $ julia
    • Activate the environment at the same location from within Julia using ] activate .
    • Include the script: include("image-gen.jl")
    • Generate the images: make_images()
  • The DEBO figure code is located in image-gen/DataPlotDEBO/
    • From the terminal, run $ python FiguresPaper.py
• Physics models for the optimization in both Python and Julia can be called from ./src/physics-models/
  • Julia
    • Start a Julia session using $ julia
    • Activate the environment at the same location from within Julia using ] activate .
      • You may need to go into the dev'd FLOWFarm repository and check out commit ec5270203786d5bcf065fff8c80bd7710906a40e
      • the version of FLOWFarm used for this work is also available at https://zenodo.org/badge/latestdoi/543173996
    • Run include("provided_model.jl")
  • Python
    • Run $ python provided_model.py