Motivation and Philosophy#
Simulating parameterized systems of equations is fundamental in science and engineering but is often computationally expensive. In many-query settings such as uncertainty quantification and optimization, reduced-order models (ROMs) are required to make these analyses tractable. Model reduction comprises a range of approaches with different tradeoffs, and no single method is universally applicable.
Projection-based reduced-order modeling (pROM) constructs reduced models by projecting the governing equations onto low-dimensional subspaces, thereby preserving key physical structure. Despite demonstrated effectiveness, pROM adoption has been limited by its intrusive implementation requirements, which constrain applicability and large-scale testing.
The Pressio ecosystem addresses this limitation by providing a framework that reduces the intrusiveness of pROMs for large-scale simulation codes and supports the development, evaluation, and comparison of reduced-order modeling methods across applications.
The Pressio EcoSystem includes:
Name |
Info |
Latest Release |
|
|---|---|---|---|
|
Suite of 1D, 2D, 3D problems spanning multiple physics and native support for sample mesh |
0.17.0 |
|
|
Header-only logging utility for Pressio libraries |
0.17.0 |
|
|
Core operations for the Pressio ecosystem |
0.17.0 |
|
|
C++ core library: ode, solvers, ROMs, etc |
0.17.0 |
|
|
Tutorials suite for the pressio C++ library |
0.17.0 |
|
|
Schwarz coupling for projection-based ROMs with Pressio |
0.17.0 |
|
|
Python bindings to the core C++ library |
0.12.0 (see disclaimer below) |
|
|
Tools and workflows for reduced-order modeling |
0.2.0 |
Warning
Disclaimer: due to limited resources/time, we are currently forced to
pause the development of pressio4py, so it will stay out of sync until
we will find the time to update it and push another release.