1
$\begingroup$

Suppose $C_1$ and $C_2$ are some fixed $n \times n$ matrices. Define the norm $\| M \| = \sum_{i = 1}^n \max_j |M_{ij}|$. What is $\min_U \|C_1 U C_2 \|$? Here $U$ ranges over the $n \times n$ unitary matrices. I would also be interested to know what is the minimizer when the norm is one of the "standard" norms, such as the Frobenius norm or the nuclear norm.

$\endgroup$
2
$\begingroup$

I can provide the answer in the case when the norm is the Frobenius norm $\|\cdot\|_{\text{F}}$. In this case, $$\|C_1UC_2\|_{\text{F}} = \sqrt{\mathrm{tr}\big((C_1UC_2)(C_1UC_2)^*\big)} = \sqrt{\mathrm{tr}(UC_2C_2^*U^*C_1^*C_1)}.$$ It is a standard result of matrix theory that minimizing this quantity over unitary matrices $U$ gives $$\mathrm{min}_U\|C_1UC_2\|_{\text{F}} = \sqrt{\sum_{j=1}^n\lambda_j\mu_{n-j+1}},$$ where $\lambda_1 \geq \lambda_2 \geq \cdots \geq \lambda_n$ are the eigenvalues of $C_1^*C_1$ and $\mu_1 \geq \mu_2 \geq \cdots \geq \mu_n$ are the eigenvalues of $C_2C_2^*$ (this follows immediately from Problem III.6.14 in Bhatia's Matrix Analysis book, for example). Furthermore, a unitary $U$ that achieves this minimum is one that makes $UC_2C_2^*U^*$ diagonal in the same basis as $C_1^*C_1$.

$\endgroup$
  • $\begingroup$ Hi! Thanks for your answer. This looks promising! $\endgroup$ – Gautam Nov 13 '19 at 21:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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