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 $(X,\mathcal{M})$ a measurable set with measure $\mu$. Let $f$ be an integrable non negative function, such that $K:=\int_{E}f \mathrm d\mu<\infty$, where $E\in(X,\mathcal M)$. Let $\alpha>0$. Calculate the following integral

$\displaystyle \lim_{n\rightarrow\infty}\int_{E} n \ln\left(1+\left(\frac{f}{n}\right)^{\alpha}\right)\mathrm{d}\mu$

I prove that for $\alpha=1$ the previous integral is $K$, and $\forall \alpha>1$ is zero. It follows by the identity $(1+x^{\alpha})\leq\alpha x$, $\forall x\geq0$ $\forall \alpha\geq1$. And the Dominated Convergence Theorem. But I don´t know how to take the case $0<\alpha<1$. Thanks in advance!

Suppose that $\alpha>1$. We take the sequece $\{f_{n}\}_{n\in\mathbb N}$ of measurable functions, $f_{n}:=n\ln(1+(f/n)^{\alpha})$. Moreover by the identity, for $\alpha>1$, we have the following $|f_{n}|=\left|n\log\left( 1+\left(\frac{f}{n}\right)^{\alpha}\right) \right|\leq n \alpha\frac{f}{n}=\alpha f$, $\alpha f$ is measurable. We need only to prove that$\{f_{n}\}_{n\in\mathbb N}$ converges pointwise to a function.

In this case $f\equiv 0$ $\forall x\in Dom(f)$, $\displaystyle\lim_{n\rightarrow \infty}n\log\left(\left(\frac{f}{n}\right)^{\alpha}+1\right)= \lim_{t\rightarrow 0}\frac{\log((ft)^{\alpha}+1)}{t}$. Making the change of variables $t=\frac{1}{n}$. And by L'Hopital, it arises $\lim_{t\rightarrow 0}\alpha f^{\alpha} \frac{t^{\alpha-1}}{(ft)^{\alpha}+1}=0$. ($\alpha-1>0$). Finally, using the Dominated Convergence Theorem, it leads to $ \displaystyle\lim_{n\rightarrow \infty} \int f_{n} \mathrm d \mu=\int 0 \mathrm d\mu =0$

share|cite|improve this question
Could you post your proof that it is zero for any $\alpha>1$? I suspect that rearranging it one should be able to prove that the integral actually diverges for any $0<\alpha<1$. – Giovanni De Gaetano May 13 '13 at 8:47
up vote 3 down vote accepted

By a familiar property of $e$ and the exponential function, we have: $$ \lim_{n\to\infty} n^\alpha \ln\left(1 + \left(\frac{f}{n}\right)^\alpha\right) = f^\alpha $$

Hence: $$ \lim_{n\to\infty} n \ln\left(1 + \left(\frac{f}{n}\right)^\alpha\right) = f^\alpha \lim_{n\to\infty} n^{1-\alpha} $$

When $0 < \alpha < 1$ and $f \ne 0$, the the limit diverges to $\infty$.

Apply Fatou's lemma: $$ \int \liminf_{n\to\infty} n \ln\left(1 + \left(\frac{f}{n}\right)^\alpha\right) \,d\mu \le \liminf_{n\to\infty} \int n \ln\left(1 + \left(\frac{f}{n}\right)^\alpha\right) \,d\mu $$

Assuming $f$ is not zero a.e., the limit on the LHS diverges to $\infty$ for a set of positive measure. It follows that the limit on the RHS is $\infty$.

share|cite|improve this answer
Thanks!! Ayman Hourieh – Aleph May 13 '13 at 9:02
@Aleph You're welcome! – Ayman Hourieh May 13 '13 at 9:02

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.