Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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 $\mu$ be a finite measure on $X \subseteq \mathbb{R}^n$.

Consider the Uniformly Integrable family $\{ f_n(\cdot) \}_{n \in \mathbb{N}}$ of functions $f_n : X \rightarrow \mathbb{R}_{\geq 0}$.

Consider a continuous, strictly-increasing function $\phi:\mathbb{R}_{\geq 0} \rightarrow \mathbb{R}_{\geq 0}$, with $\phi(0)=0$ such that $\phi(f_n(\cdot))$ is integrable for all $n$.

  1. Is the family $\{ \phi( f_n(\cdot) ) \}_{n \in \mathbb{N}}$ Uniformly Integrable as well?

  2. More generally, with functions $g_n : X \rightarrow \mathbb{R}^n$ such that the family $\{ |g_n(\cdot)| \}_{n \in \mathbb{N}}$ is Uniformly Integrable, and a continuous $\psi: \mathbb{R}^n \rightarrow \mathbb{R}_{\geq 0}$, is $\{ \psi(g_n(\cdot)) \}_{n \in \mathbb{N}}$ Uniformly Integrable?

share|cite|improve this question
no, they are not even necessarily integrable,e,g,$\phi(x) = x^2$ – mike Jun 6 '12 at 23:44
Why? Counterexample? If it is not the case, under which conditions we have that UI is preserved? – Adam Jun 6 '12 at 23:52
I see your comment. Of course, this is a trivial condition that has to be satisfied. I'll update the question accordingly. – Adam Jun 7 '12 at 0:00
up vote 3 down vote accepted

Consider $n=1$ and $f_j=j\chi_{(0,\frac 1{j^2})}$. Fix $R>0$, when $j\geq R$ we have $$\int_{|f_j|\geq R}|f_j|dP\leq \sqrt{\mu(|f_j|\geq R)}\leq \frac 1R,$$ hence $\{f_j\}$ are uniformly integrable, and in $L^3$. We have $\lVert f_j\rVert_{L^3}=j\left(\int_{(0,\frac 1{j^2})}1\right)^{1/3}=j^{1/3}$, hence the family $\{f_n^3\}$ is not uniformly integrable.

share|cite|improve this answer
What additional assumptions should be made on $\phi$ to get that $\{ \phi(f_n(\cdot)) \}_n$ are UI? – Adam Apr 2 '13 at 9:17
Moreover, in your example, are the functions $\phi(f_n(\cdot))$ integrable? – Adam Apr 2 '13 at 9:23
Yes, $\phi(f_n)$ is bounded and the measure space is finite. For conditions about preservation of uniform integrability, I presume you already know de la Vallée Poussin's theorem, so it could be a start. – Davide Giraudo Apr 2 '13 at 9:24
Yes, I do know that and I am not clear about your comment. In the de la Vallée Poussion Theorem, we have a scalar function $G$ such that $\lim_{x \rightarrow \infty} G(x)/x = \infty$. Here we want to use like $|f|$ is place of $G$, but it depends on the parameter $n$. – Adam Apr 2 '13 at 9:37

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.