Singular Value Decomposition of Commuting Matrices

599 Views Asked by At

If two square matrices $M_1$ and $M_2$ commute, does it mean that the $U$ and $V^\dagger$ appearing in their singular value decompositions are the same?

Specifically, does it imply that

$$M_1 = U \Sigma_1 V^\dagger$$ $$M_2 = U \Sigma_2 V^\dagger$$

for two possibly different $\Sigma_{1,2}$?

As $M_1$ and $M_2$ commute, I know that they can be simultaneously diagonalized, which means there's an orthogonal transformation $\mathcal{O}$ which can diagonalize both matrices. But does this have any bearing on $U$ and $V$ of the singular value decomposition?

1

There are 1 best solutions below

1
On BEST ANSWER

Counter-example: take $M_1 = I$, the identity matrix, and $M_2$ as any matrix of the same dimension (not a scalar multiple of the identity). You can see the above statement fails to hold.