I was going through the proof of the Dominated Convergence Theorem.

Now if we have that ($f$$_n$) is a sequence of measurable functions such that $\lvert f_n\rvert$ $\le$ $g$ for all n where g is integrable on $\Bbb{R}$. And if $f$ = $\lim_{n}$$f_n$ almost everwhere.

We can show that ($g+f_n$) is a sequence of non-negative measurable functions.

Then by Fatou's lemma, we have that $\int$liminf($g+f_n$)$d$$x$ $\le$ liminf$\int$($g+f_n$)$d$$x$.

Now from here, we can obtain that

$\int$($g+f$)$d$$x$ $\le$ $\int$$gdx$+liminf $\int$$f_ndx$.How?

Please explain this last step.I know that since both are integrable, the integral can be seperated..but how is liminf seperated in the right-hand side?


Inside the liminf, $\int g$ is just a number. So what it says is that $$ \liminf_n(c+f_n)=c+\liminf _nf_n, $$ where $c=\int g$.

  • $\begingroup$ So..is that always true? ..liminf(c+an) = c+liminfan? $\endgroup$
    – Gitika
    Jun 10 '20 at 5:19
  • $\begingroup$ Yes. How could translating a set by a fixed distance could change it in any other way? If you move a set 10 units to the right, the infimum will increase by 10. $\endgroup$ Jun 10 '20 at 5:21
  • $\begingroup$ Are you sure $\int$g is a number? $\endgroup$
    – Gitika
    Jun 10 '20 at 5:39
  • $\begingroup$ for example..what if g(x)=x? $\endgroup$
    – Gitika
    Jun 10 '20 at 5:40
  • $\begingroup$ I can see it is independent of n..it's a constant sequence $\endgroup$
    – Gitika
    Jun 10 '20 at 5:41

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.