DPDL
From a description of a physical system to a computational model.
View source ↗
The question
Can a concise description of a physical system become the starting point for its solver, instead of rewriting the numerical workflow for every model?
The approach
The repository documents a Python decorator for Lagrangian expressions, symbolic derivation with SymPy, and a physics-informed neural-network solver using PyTorch.
Explore the work
Read the examples and implementation in the repository. This portfolio describes the documented approach; it does not run training or present synthetic compiler output as a result.