MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $f_\alpha$ be a family of continuous positive functions $\mathbb R\to \mathbb R$ where the index $\alpha$ runs in a compact metric space and the map $\alpha\to f_\alpha$ is continuous with respect to compact-open topology on the target. Suppose there is a uniform upper bound on the integrals of $f_\alpha$'s over $\mathbb R$.

Question. Is $\underset{\alpha}{\sup} f_\alpha$ necessarily an integrable function?

Apology. This sure sounds like a homework level question, but after looking at for a while I am not even sure what the answer is.

share|cite|improve this question
Note that in your hypotheses the family $\{f_\alpha\}$ is equicontinuous on bounded intervals, so the supremum is a continuous function (possibly non integrable). – Pietro Majer May 16 '11 at 16:48
Also, every nonnegative continuous function $f\colon\mathbb{R}\to\mathbb{R}$ can occur as $\sup\_\alpha f_\alpha$. Consider the case where the metric space is the extended reals $\mathbb{\bar R}$, $f_\alpha=0$ for $\alpha=\pm\infty$ and $f_\alpha(x)=f(x)(1-g(\alpha)\vert\alpha-x\vert)\_+$ otherwise, with $g\colon\mathbb{R}\to\mathbb{R}$ a continuous function tending to inifinity as $\vert\alpha\vert\to\infty$ fast enough that $\int f_\alpha(x)\,dx$ is bounded. – George Lowther May 16 '11 at 17:21
up vote 10 down vote accepted

No. Let the compact metric space be $[0,1]$ with the standard topology and define $$ f_\alpha(x) =\alpha\max(1-\alpha \vert x\vert,0) $$ for $\alpha\in[0,1]$. This satisfies the properties asked for, with the upper bound $\int f_\alpha(x)\\,dx\le1$ (and, equality for $\alpha\not=0$). But, $$ \sup\_\alpha f_\alpha(x)=1_{\{\vert x\vert\ge1/2\}}\frac{1}{4\vert x\vert}+1_{\{\vert x\vert < 1/2\}}(1-\vert x\vert) $$ is not integrable.

share|cite|improve this answer
Thanks so much! – Igor Belegradek May 16 '11 at 16:48

As you stated, no: take e.g. the unit closed interval $[0,1]$ as a compact metric space and define for all $\alpha\in [0,1]$ the function $f_\alpha(x)= \alpha (1-\alpha|x|) _ +$, that depends continuously on $\alpha$ even in the uniform norm on $\mathbb{R}$. Then for $|x|\ge 1$ we have $ \sup_ {\alpha\in[0,1]} f_ \alpha (x)=1/4x $, which is not integrable.

share|cite|improve this answer
Thanks so much! I accepted George Lowther's answer because he was there first. :) – Igor Belegradek May 16 '11 at 16:48

Just one more example: let $\phi \in C_c(\mathbb{R})$ with, say, $\sup_{\mathbb{R}} \phi = 1$. For $\alpha \in [-\frac{\pi}{2}, \frac{\pi}{2}]$, set $$f_\alpha(x) = \begin{cases} \phi(x - \tan \alpha), & \alpha \in (-\frac{\pi}{2}, \frac{\pi}{2}) \\\\ 0, &\alpha = \pm \frac{\pi}{2}. \end{cases}$$ Then $\int f_\alpha = \int \phi < \infty$ for each $\alpha \ne \pm \frac{\pi}{2}$, and $\int f_\alpha = 0$ otherwise, but $\sup_\alpha f_\alpha = 1$.

share|cite|improve this answer
Very nice, thank you! – Igor Belegradek May 16 '11 at 23:05

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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