RBmatlab is a MATLAB library for model order reduction with Reduced Basis Methods
for various discretization types and application settings.
Firstly the RBmatlab library provides discretization techniques and meshes to implement simple models. In case that more complex models are to be treated, it provides interfaces to external packages as DUNE or ALBERTA-Grid. Some of the basic features concerning discretization are:
- Grid management (2D rectangular, triangular, Multi-D adaptive nonconforming cubegrid).
- Discretization techniques: Finite Elements, Finite Volumes, Finite Differences, Local Discontinuous Galerkin including discrete function types.
- Visualization.
Secondly the library provides model reduction algorithms for Reduced Basis model reduction of parameter dependent stationary and instationary problems. Some of the key features concerning model order reduction are:
- (POD-) Greedy algorithm for basis generation.
- Reduced basis simulations and error estimation for various kinds of problems.
- Treatment of nonlinearities and non-affine parameter dependence via Empirical Operator Interpolation.
- Basis generation including hp, p-partition, t-partition algorithms.
Furthermore the library provides a variety of demo files and implemented problems (linear advection-diffusion, two phase flow,...).
On the website morepas.org one finds a detailed documentation as well as information how to download the software. Please note: The RBmatlab package contains several precomputed data files for demonstrations and an extensive html documentation. Therefore the total size of the package is approximately 100MB (zipped) or 150 MB (unzipped).