2
$\begingroup$

Let $m\in C[a,b]$. Consider on $(C[a,b], \|\cdot \|_{\infty})$ the multiplication operator $A: C[a,b] \to C[a,b], \quad Af = mf$.

Prove that $\|A\| = \|m\|_{\infty}$.

In my book, we are given the following definition although it is written as an example:

A bounded sequence $\lambda = (\lambda_n)_{n \in \mathbb{N}}$ induces a multiplication operator $A_{\lambda} : \ell^2 \to \ell^2$ by $(A_{\lambda}f)(n):= \lambda_nf(n), \quad n \in \mathbb{N}, f \in \ell^2$.

Furthermore, the author provides a brief proof for elements of $\ell^2$ which I have modified to fit my assumptions which is shown below:

Since $\lambda$ is a bounded sequence, we have

$\|A_{\lambda}f\|_{\infty} = \sup\{\int_{a}^{b}|A_{\lambda}f(t)|dt\} \leq \sup\{\int_a^b \|A_{\lambda}\|_{\infty} |f(t)|dt\} = \|A_{\lambda}\|_{\infty} \|f\|_{\infty}$ for all $f \in C[a,b]$. So $A_{\lambda}$ is bounded and $\|A_{\lambda}\| \leq \|\lambda\|_{\infty}$.

On the other hand, $\|A_\lambda e_n\|_{\infty} = \|\lambda_n e_n\|_{\infty} = |\lambda_n| \|e_n\|_{\infty} = |\lambda_n|$ where $e_n$ is the $n$th standard unit vector. Hence, $\|A_{\lambda}\| \geq |\lambda_n|$ for every $n \in \mathbb{N}$, and thus $\|A_{\lambda}\| \geq \sup_n |\lambda_n| = \|\lambda\|_{\infty}$. Using both these estimates gives $\|A_m\| = \|m\|_{\infty}$.

Can somebody please explain to me if this is a proper imitation of the proof in $\ell^2$ for the supremum norm?

Thanks in advance, your assistance is greatly appreciated.

My textbook is Functional Analysis An Elementary Introduction by Haase.

$\endgroup$
  • $\begingroup$ I'm somewhat confused: if you were trying to prove $\|A_\lambda\| = \|\lambda\|_\infty$, why did you mention the original problem ($\|A\| = \|m\|_\infty$ for $A : C[a,b] \to C[a,b]$, $Af = mf$)? $\endgroup$ – kobe Feb 17 '15 at 17:54
  • $\begingroup$ I had seen a very similar solution involving the $\ell^2$ space and was trying to use that as a model for my problem. $\endgroup$ – Jamil_V Feb 17 '15 at 18:53
2
$\begingroup$

By the definition of operator norm, $$ \| A\|=\sup\{ \| Af\|_{\infty};\quad \| f\|_\infty \leq 1\}. $$ Note that $|m(t)f(t)|\leq \| m\|_\infty \| f\|_\infty$ $(\forall t\in [a,b])$. Hence $$ \| Af\|_\infty= \| m f\|_\infty\leq \| m\|_\infty \| f\|_\infty. $$ It follows that $$ \| A\|=\sup\{ \| Af\|_{\infty};\quad \| f\|_\infty \leq 1\}\leq \|m\|_\infty. $$ On the other hand, for the constant function $e(t)=1$ we have $\| e\|_\infty=1$ and $$ \| A e\|_\infty=\| m e\|_\infty=\| m\|_\infty. $$ Thus, $\| A\|=\| m\|_\infty.$

$\endgroup$
  • $\begingroup$ Thank you very much for clarifying, I see now exactly where my confusion was! $\endgroup$ – Jamil_V Feb 17 '15 at 20:15

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.