Shape optimization is a natural application of many of the ideas which went into building the Tesseract framework. We wrote a demo where we solve an optimization problem to find structural configurations that minimize compliance using gradient-based optimization through JAX-FEM and geometry processing.
The pipeline involves two Tesseracts with computation orchestrated via Tesseract-JAX; it’s interesting to note how this makes composing Tesseracts look exactly like composing local Python functions, and also how libraries like optax can be used in conjunction with
Check out the demo here
