I am facing following equation:
$$ A * X + C \cdot X = D $$
with:
- $A, C, D \in \mathbb{R}^{n \times n}$ some known matrices without any particular structure,
- $X \in \mathbb{R}^{n \times n}$ the unknown matrix
- "$\cdot$" the element-wise multiplication
- "$*$" the cyclic convolution defined as: $A * X = \mathcal{F}^{-1}( \mathcal{F}(A) \cdot \mathcal{F}(X) ) $.
Is there a way to solve this? An analytic way would be fantastic but fast numerical methods are also welcome (for instance, I thought of minimizing the error... but not very elegant nor efficient).
I already tried to make some Fourier transforms to get rid of the convolution but unsuccessfully so far... Any help will be highly appreciated!!!