Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $A,B\in\mathbb{C}^{n\times 3n}$ be two matrices, and denote the Kronecker matrix product by $\otimes$. The matrix $$ M= \begin{bmatrix} A \otimes I_n \\\\ I_n \otimes B\end{bmatrix} $$ has size $2n^2 \times 3n^2$, therefore in the generic case its kernel has dimension $n^2$.

Is there a simple characterization of this kernel, or a way to compute it by exploiting this structure?

This problem comes from a numerical linear algebra problem (two-parameter eigenvalues problems). Do matrices with this structure appear in other fields, or do they ring a bell to you?

It looks appealing to look for vectors in this kernel in the form $v=a \otimes b \otimes c$, with $a,c\in\mathbb{C}^{n}$, $b\in\mathbb{C}^{3}$, since in this way they would be compatible with both Kronecker product structures. We normalize by restricting one entry of each of $a,b,c$ to be $1$, so we are left with $(n-1) + 2 + (n-1) = 2n$ unknowns. The condition $Mv=0$ is equivalent to $M(a\otimes b)=0$ and $N(b\otimes c)=0$, which are precisely $2n$ equations. So it looks like we should have several complex solutions in the generic case. Is there a viable way to compute them?

share|improve this question
    
Isn't the relation $vec(ABC) = (C^T \otimes A) vec(B)$ helpful? –  Dirk Jan 24 '13 at 22:00
    
I tried playing with that relation, but I don't see how to put it to good use. Note that the block sizes in the two $\otimes$'s are different, so you can't use it to rewrite the two block rows as matrix products with the same format. –  Federico Poloni Jan 24 '13 at 22:18
add comment

1 Answer 1

Not an answer but too many equations for a comment: $\newcommand{\vec}{\mathrm{vec}}$ To compute an element $v\in\mathbb{R}^{3n^2}$ of the kernel of $$M = \begin{bmatrix} A\otimes I\\\\ I\otimes B\end{bmatrix}$$ we use the formula $\vec(ABC) = (C^T\otimes A)\vec(B)$ which holds whenever the products are defined.

Now we can write $v\in\mathbb{R}^{3n^2}$ as both $v=\vec(V)$ with $V\in\mathbb{R}^{n\times 3n}$ (stacking the $3n$ column vectors of length $n$ on top of each other) and $v= \vec(U)$ with $U\in\mathbb{R}^{3n\times n}$ (stacking the $n$ column vectors of length $3n$ on top of each other).

We write $$Mv = \begin{bmatrix} (A\otimes I)\vec(V)\\\\ (I\otimes B)\vec(U)\end{bmatrix} = \begin{bmatrix} \vec(VA^T)\\\\ \vec(BU)\end{bmatrix}.$$ Hence, the vectors $v=\vec(V)=\vec(U)$ in the kernel of $M$ are characterized by $$AV^T = 0,\qquad BU = 0.$$ Note that $V$ and $U$ are related by a reshape of their elements. Looks like this could be exloited algorithmically but I did not see yet how...

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.