Installation and requirements

This page describes how to install the dependencies used by CFDNNetAdapt.

What the repository contains

CFDNNetAdapt combines multi-objective evolutionary optimization with deep-neural-network surrogate models. The repository contains:

• the framework implementation in src/;
• local copies of the Platypus MOEA and pyrenn code in thirdParty/.

Requirements

Common requirements

• Python 3.10 or newer;
• Python packages:
  • numpy
  • scipy
  • dill
  • matplotlib (optional visualization of results);

Platypus and pyrenn are included locally under thirdParty/; they do not need to be installed separately from PyPI.

OpenFOAM requirements

The CFD example additionally requires:

• OpenFOAM v8;
• an OpenFOAM environment script available at /opt/openfoam8/etc/bashrc, as referenced by the supplied run scripts.

If OpenFOAM is installed elsewhere, update the source line in Allrun-serialAll and Allrun-test.

Installation

Clone the repository and create an isolated Python environment:

git clone https://github.com/LucieKubickova/CFDNNetAdapt.git
cd CFDNNetAdapt
python3 -m venv .venv
source .venv/bin/activate
python3 -m pip install --upgrade pip
python3 -m pip install numpy scipy dill matplotlib

Confirm that the Python dependencies and local modules can be imported:

python3 - <<'PY'
import dill
import matplotlib
import numpy
import scipy
import sys

sys.path.insert(0, "src")
sys.path.insert(0, "thirdParty")
import CFDNNetAdaptV3
import pyrennModV3
print("CFDNNetAdapt dependencies imported successfully")
PY

No requirements.txt or packaging configuration is currently provided, so the explicit pip install command above is the supported setup method.