Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have two diagonal matrices over the reals, both contain positive values on the diagonal (and 0 everywhere else), $A$ and $B$. $A$ is of size $n \times n$ and $B$ is of size $m \times m$. We assume that $n > m$.

I would like to solve the equation for $X$:

$X A X^{\top} = B$

where $X$ is an $m \times n$ matrix.

Clearly, one solution would be to set $X = B^{1/2} C$ where $C$ of size $m \times n$ with $C_{ii} = $A_{ii}^{-1/2} and 0 everywhere else.

My question - is the solution for $X$ unique (and it is the one I described)? If not, to what extent is it unique?

EDIT: I just realized that the $A$ matrix is a red-herring. One could define $C$ as above, and then substitute $Y = X C$ (which is an invertible linear transformation) to get the equation:

$Y Y^{\top} = B$.

So the question is, if $n > m$, is the solution for $Y \in \mathbb{R}^{m \times n}$ unique for the above equation when $B$ is diagonal, with positive values over the diagonal? (up to the sign of $Y$, because if $Y$ is a solution, then so is $-Y$.)

share|improve this question
1  
+1 For the edit. –  Alex Becker May 9 '12 at 22:24
1  
Thanks Alex. Maybe I was too quick to send this question - I think the answer is "no". One could have Y be just scaled orthonormal vectors. –  kloop May 9 '12 at 22:28

1 Answer 1

up vote 3 down vote accepted

The answer is definitely "no". Given any solution $Y$ and an orthogonal $n\times n$ matrix $U$, then $YU$ is also a solution.

share|improve this answer

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.