Archimedes is an open-source Python framework designed for deployment of control systems to hardware. To make this possible, it provides a comprehensive toolkit for modeling, simulation, optimization, and C code generation. Archimedes builds on the powerful symbolic computation capabilities of CasADi with an interface designed to be familiar to NumPy users.
Introduction to the Archimedes project
Learn how to use Archimedes
Announcements, deep dives, and case studies
Quick Example¶
See the Quickstart for recommended installation; the easiest version is:
pip install archimedes
Archimedes combines design features from SciPy, JAX, and PyTorch to make it possible to write regular NumPy functions that can be automatically differentiated and translated to C code for embedded applications:
import numpy as np
import archimedes as arc
def f(x):
return np.sin(x**2)
# Use automatic differentiation to compute df/dx
df = arc.grad(f)
np.allclose(df(1.0), 2.0 * np.cos(1.0)) # True
# Automatically generate C code
template_args = (0.0,) # Data type for C function
arc.codegen(df, "grad_f.c", template_args, return_names=("z",))
Key features:
NumPy-compatible array API with automatic dispatch
Efficient execution of computational graphs in compiled C++
Automatic differentiation with forward- and reverse-mode sparse autodiff
Interface to “plugin” solvers for ODE/DAEs, root-finding, and nonlinear programming
Automated C code generation for embedded applications
JAX-style function transformations
PyTorch-style hierarchical data structures for parameters and dynamics modeling
Documentation¶
Resources
Core Concepts
Applications
Reference
Development