0
$\begingroup$

Show that $$\|AC\|_2 \leq \|A\|_2 \|C\|_2,$$

where $$A, C$$ are square matrices of the same size.


I understand this probably needs to use Cauchy-Schwarz Inequality, but I don't know how should I start the problem, any help?

$\endgroup$
2
$\begingroup$

Observe that $$ \|AC\|=\sup_{\|x\|=1}\|ACx\|\le \sup_{\|x\|=1}\|A\|\|Cx\|=\|A\|\sup_{\|x\|=1}\|Cx\|=\|A\|\|C\|. $$ Here we only used that $$ \|ACx\|\le \|A\|\|Cx\| $$ which is an implication of the definition of $\|A\|$.

Note. This inequality HAS NOTHING TO DO with Cauchy-Schwarz, although it looks the same!

$\endgroup$
2
$\begingroup$

The answer will depend on your definition of $\|\cdot \|_2$.

In the case of the induced $2$-norm. The answer is given by Yiorgos S. Smyrlis.

In the case of the "entrywise" $2$-norm, we see that \begin{align} \|AC\|_2^2 =&\ \sum^n_{j=1}\sum^n_{i=1}|(AC)_{ij}|^2 =\sum^n_{j=1}\sum^n_{i=1}\left|\sum_{k=1}^n a_{ik}c_{kj} \right|^2\\ \leq&\ \sum^n_{j=1}\sum^n_{i=1} \left(\sum^n_{k=1}|a_{ik}|^2 \right)\left(\sum^n_{k=1}|c_{kj}|^2 \right)\\ =&\ \left(\sum^n_{i=1}\sum^n_{k=1}|a_{ik}|^2\right)\left(\sum^n_{j=1}\sum^n_{k=1}|c_{kj}|^2\right) = \|A\|_2^2\|C\|_2^2 \end{align} which we used Cauchy-Schwarz.

$\endgroup$

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.