This repo contains a python script for automatically detecting "stable" chaotic systems. The script can search for them, store their initial configuration and later displaying the system after N iterations in an artistic style.

git clone [email protected]:AIRLegend/attractors.git
cd attractors
pip install -r requirements.txt

python attractors.py search --n_searches 1500 --search_out_path ./sols/

This command will spawn the search of 1500 chaotic systems and save the initial configurations next to a simple plot of the result in the sols/ directory.

After having generated "solution files" you can use them for creating their visualizations.

python search.py display --solution_file ./sols/sol_0_1715774254.128875.json --cmap Blues --figsize 30 30 --alpha 0.7 --n_samples=200000

You can also generate an animation with the video parameter:

python attractors.py video --solution_file ./sols/sol_10_1715885863.353617.json --cmap cividis --figsize 20 20

(See help for the extra supported styling options)

You can generate visualizations such as:

I've found these resources very interesting. Worth checking out if you want to learn more about this stuff!

• https://paulbourke.net/fractals/lyapunov/
• https://en.wikipedia.org/wiki/Lyapunov_exponent
• Automatic Generation of Strange Attractors, Sprott, J. 1993
• https://en.wikipedia.org/wiki/Attractor