Skip to main content
All experiments
Generative
Live

Strange Attractors

Lorenz, Rossler, and Chen attractor systems rendered as SVG paths for pen plotting. Part of the broader pen-plotter autoresearch loop.

Topographic line study from the pen plotter autoresearch catalog

Sample line study from the autoresearch catalog · view the field journal →

The three classical strange attractors, Lorenz (sigma=10, rho=28, beta=8/3), Rossler, and Chen, are solved numerically using a 4th-order Runge-Kutta integrator. The solver runs with configurable dt and num_steps, producing X/Y/Z coordinate traces that are then projected into 2D and serialized as SVG path data.

Output is optimized for physical pen plotting on an iDraw. This means path ordering matters: the generator clusters nearby strokes to minimize pen-up travel, and outputs a single continuous path where possible to avoid unnecessary lifts. Line width is modulated by the local velocity of the attractor trajectory, giving the strokes a natural weight variation.

The attractor generators feed into the larger pen-plotter autoresearch corpus alongside flow fields, op-art, and constructivist studies. The full scored catalog with 25,040 specimens is browsable in the field journal.

Highlights

  • 4th-order Runge-Kutta solver with configurable dt and step count
  • SVG path output optimized for pen plotting: minimal pen lifts, velocity-weighted stroke width
  • All three classical attractors: Lorenz, Rossler, Chen
  • Outputs feed the pen plotter autoresearch catalog alongside 21 other factories

Stack

  • Python
  • NumPy
  • SVG
  • matplotlib
  • iDraw