6
$\begingroup$

I know that functions in $L^2$ space have finite norms by definition, but are they also bounded "almost everywhere" ?

So say for instance the following functions norm is finite but it is not bounded.

$$ f(x) = \frac{1}{(x-\frac{1}{2})^2} ~~ ; ~~~ f(x) \in L^2[(0,1),\mu] $$

$$ {\|f(x)\|}_2 = \left(\int_0^1{\frac{1}{(x-\frac{1}{2})^2}}\right)^{1/2}$$ $$ {\|f(x)\|}_2 = \left(\left. {\frac{1}{(x-\frac{1}{2})}}\right]_0^1\right)^{1/2}$$ $$ {\|f(x)\|}_2 = 2$$

$\endgroup$
  • $\begingroup$ Where is the $p$ in the integral? Are you already answering the question in the question? $\endgroup$ – Sergio Parreiras Dec 13 '13 at 5:48
  • $\begingroup$ I meant in general fo L^p space, but I have tried to give a specific example. $\endgroup$ – Michael T Mckeon Dec 13 '13 at 5:51
  • 1
    $\begingroup$ Michael, functions in $L^p$ spaces are equivalence classes of functions - you can change a function's values on a zero measure set and it stays in the same equivalence class. So you only need to question whether your function is bounded almost everywhere. $\endgroup$ – massy255 Dec 13 '13 at 5:51
  • $\begingroup$ OK, ill edit the question, I understand your point. $\endgroup$ – Michael T Mckeon Dec 13 '13 at 6:08
  • $\begingroup$ see this related post: math.stackexchange.com/questions/458437/… $\endgroup$ – massy255 Dec 13 '13 at 6:15
11
$\begingroup$

Let's fix the definitions first.

A function $f:(0,1)\to\mathbb R$ is called bounded if there is a number $M$ such that $|f(x)|\le M$ for all $x\in (0,1)$.

Since the elements of $L^p$ are not functions, but equivalence classes of functions, the concept of boundedness must be adjusted.

A function $f:[0,1]\to\mathbb R$ is called essentially bounded if there is a number $M$ such that $|f(x)|\le M$ for almost all $x\in (0,1)$. (That is, the inequality holds on some set $E$ such that $(0,1)\setminus E$ has zero measure.)

If two functions $f,g$ belong to the same equivalence class and $f$ is essentially bounded, then $g$ is also essentially bounded. Thus, the property of being essentially bounded makes sense for elements of $L^p$.

Does it hold for all elements of $L^p$? Yes if $p=\infty$ (by definition), no if $p<\infty$. For example, the function $f(x)=x^{-1/(p+1)}$ defines an element of $L^p((0,1))$ which is not essentially bounded. For every $M$, the set $\{x:|f(x)|>M\}$ contains an interval, and therefore is not a null set.


Remark. One often allows functions to take values in the extended real line $[-\infty, \infty]$. In this case, one can say that every element of $L^p$ is finite almost everywhere.

$\endgroup$
  • $\begingroup$ Concerning your remark, isn't any element of Lp(I) finite almost everywhere, whatever the interval I? Not only for the extended real line? $\endgroup$ – Stephane Jul 23 at 16:03

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.