1
$\begingroup$

Let $C[0, 1]$ be the linear space of all continuous functions on the interval $[0, 1]$ equipped with the norm $\|f\| = \max_{0≤x≤1} |f(x)|$. Let $K(x, y)$ be a fixed function of two variables, continuous on the square $[0, 1] \times [0, 1]$, and let $A$ be the operator defined by $$Af(x) = \int^{ 1}_0 K(x, y)f(y) dy.$$ For $K(x, y) = x$, find $\|A\|$.

Proof idea:

$A$ is a continuous linear operator mapping $C[0,1]$ on itself. Since $A$ is linear from linearity of integrals. Also $A$ is bounded since $K$ is continuous on a compact set $[0,1]\times[0,1]$ and thus bounded by some $M$. Therefore, $$|Af(x)|=\left|\int_0^1K(x,y)f(y)dy\right|\leq \int_0^1|K(x,y)||f(y)|dy\leq\int_0^1M\|f\|=M\|f\|.$$

Now, if I am correct about the above, $|Af(x)|\leq x\|f\|$ for $K(x,y)=x$.

$$\|A\|=\sup_{f(x)\neq0}\frac{\|Af(x)\|}{\|f(x)\|}=\sup_{f(x)\neq0}\frac{\|Af(x)\|}{\|f(x)\|}\leq\|f\|?$$ is $\|A\|=\sup|Af(x)|$? Is the final answer an inequality? I am confused about this.

$\endgroup$
3
  • 1
    $\begingroup$ The final answer is a specific positive real number, a supremum of a bounded set. Since $1$ bounds $K$, you also already know that this supremum is $\le 1$. If you find a function (or a sequence of functions) of norm $1$ that achieves (converges to) $1$, then you proved $\|A\|=1$. $\endgroup$
    – Berci
    Nov 11 '21 at 19:09
  • $\begingroup$ Is this part two of a two-part problem, with the first part being "Show $A$ is bounded"? You have a lot of extra work prior to "Now, if I am correct about the above", and then $|Af(x)| \leq x \|f\|$, while true, is currently unjustified. $\endgroup$ Nov 11 '21 at 19:25
  • $\begingroup$ Yes, you are correct @BrianMoehring. Part a was: Prove that A is a continuous linear operator mapping C[0, 1] into itself. $\endgroup$
    – Mihai.Mehe
    Nov 11 '21 at 19:26
1
$\begingroup$

Is $\|A\|=\sup|Af(x)|$?

This isn't even clear what you mean by the supremum, but I'm apt to say no*. The definition is $$\|A\| = \sup_{\|f\|=1} \|Af\| = \sup_{f\neq 0} \frac{\|Af\|}{\|f\|}$$ Since you also have the norm $$\|Af\| = \max_{0 \leq x \leq 1} |Af(x)| = \sup_{0 \leq x \leq 1} |Af(x)|$$ floating around, you need to be more clear about your intentions when you write the supremum.

[*If you meant $\|A\| = \sup_\limits{\|f\|=1 \\ 0 \leq x \leq 1} |Af(x)|$, then the answer would be "yes", but it would be an irregular use that doesn't match anything else you've written]

Is the final answer an inequality?

If you fix a couple notational issues (and one error) in your solution, you'll have shown an inequality $\|A\| \leq 1$. However, this is not the final answer, since you are tasked with actually computing $\|A\|$.


Let's clean up the notation. I'll run through a full solution (in particular, note where I've included $x$ and where I didn't):

First, setting $K(x,y)=x$, we have $$|Af(x)| = \left|\int_0^1 xf(y)\,dy\right| \leq x\int_0^1|f(y)|\,dy \leq x\int_0^1\|f\|\,dy = x\|f\|$$ Therefore $$\|Af\| = \max_{0\leq x \leq 1}|Af(x)| \leq \max_{0\leq x \leq 1} x\|f\| = \|f\|$$ and $$\|A\| = \sup_{f \neq 0}\frac{\|Af\|}{\|f\|} \leq 1,$$

To show the other direction, we need to go back to the definition. Noting that $f_0(x) = 1$ implies $\|f_0\| = 1$ and $Af_0(x) = x$, we have $$\|A\| \geq \frac{\|Af_0\|}{\|f_0\|} = \|Af_0\| = \max_{0 \leq x \leq 1} |Af_0(x)| = \max_{0 \leq x \leq 1} |x| = 1$$ so $\|A\| = 1$.

$\endgroup$
1
$\begingroup$

By rearranging what you have written (and remembering to divide by the norm of $f$ in your expression), you have shown the inequality $\|A\|\leq 1$, and in this instance, this is actually the norm.

To see this, note that the norm is achieved by $Af$, for $f\equiv 1$.

$\endgroup$
1
  • $\begingroup$ $$\|A\|=\sup_{f(x)\neq0}\frac{\|Af(x)\|}{\|f(x)\|}\leq||f||$$ since $||Af(x)||=sup_{f(x)\leq1}|Af(x)|=sup_{f(x)=1}|Af(x)|$ and $||f(x)||=sup_{x=1}|f(x)|$ then $||A||=1$. $\endgroup$
    – Mihai.Mehe
    Nov 11 '21 at 19:42

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.