`pymor.algorithms.simplify`¶

Concatenations of non-parametric linear Operators are contracted into a single Operator when possible. Products of sparse matrices are not computed, however.
Non-parametric linear combinations of non-parametric operators are merged.

Parameters:: obj – Either a Model or an Operator to which the contraction rules are applied recursively for all children.
Returns:: The transformed object.

pymor.algorithms.simplify.expand(obj)[source]¶

Expand concatenations of LincombOperators.

To any given Operator or Model, the following transformations are applied recursively:

(O1 + O2) @ (O3 + O4)

becomes:

O1 @ O3 + O1 @ O4 + O2 @ O3 + O2 @ O4

LincombOperators inside LincombOperators are merged into a single LincombOperator.
ConcatenationOperators inside ConcatenationOperators are merged into a single ConcatenationOperator.
Adjoints of LincombOperators are distributed over the linear combination
```
(∑ c_i A_i)^* = ∑ conj(c_i) A_i^*.
```
ConcatenationOperators inside AdjointOperators are reversed
```
(A @ B @ C)^* = C^* @ B^* @ A^*.
```
BlockOperators are expanded block-wise. Further, LincombOperator blocks are expanded into a single LincombOperator of BlockOperators.

Parameters:: obj – Either a Model or an Operator to which the expansion rules are applied recursively for all children.
Returns:: The transformed object.

pymor.algorithms.simplify¶