Petar Mlinarić
Petar Mlinarić
So far, only the Gramians from `BTReductor` were cached, which was done by the `LTIModel`.
Sometimes, I find pylint useful, at least after ignoring the majority of its warnings. Are others fine with adding these settings?
SciPy 1.8 introduced [sparse arrays](https://docs.scipy.org/doc/scipy-1.8.0/html-scipyorg/reference/sparse.html#module-scipy.sparse). It might be that pyMOR already supports both sparse matrices and arrays, but it would be good to actually test this.
The goal here is to add functionality to adaptively sample the transfer function given the boundaries of the frequency interval over interest (see #389). So far, I added an `adaptive`...
There are a couple of issues which might be good to discuss: 1. RB and sys-mor parts use reductors differently (sys-mor methods consist of instantiating a reductor and calling its...
This could be an approach to computing the l2 norms of dofs in a `VectorArrays` (by squaring the elements and summing the vectors), which is necessary for [QDEIM](https://doi.org/10.1137/15M1019271) (but not...
This is necessary for QR decomposition (with column pivoting) of tall and skinny matrices using [Householder reflections](https://en.wikipedia.org/wiki/QR_decomposition#Using_Householder_reflections), where the matrix is represented using a `VectorArray` of its columns. Another application...
When running the following code (which uses [perfplot](https://github.com/nschloe/perfplot)) ```python import numpy as np import numpy.linalg as npla import scipy.linalg as spla import perfplot # setup n = 5000 rng =...
This PR is similar to #1054 but is limited to `solve` and `output`. I would be for adding impulse response and step response once #1305 is merged, as I believe...
For now, this is a draft PR implementing methods mentioned in #389, which might help with discussing how to handle input functions. Currently, `LTIModel._solve` expects an `inputs` keyword argument with...