I want to show that if $f_n \rightarrow f$ a.e. on $A, \lambda(A) < \infty$ then $\lim_{n\rightarrow \infty} \int_A f_n(x) dx = \int_A f(x) dx$.

I've tried to prove this but my proof is sitting uncomfortably with me and I think it's missing things. In addition the question also has the definition of an equiintegrable sequence of functions in it, which I haven't used. If anyone could point out what I've done wrong or what I'm missing I'd be really grateful! Thanks in advance!

What I've done: Using Egorov's theorem since $\lambda(A)<\infty$ and $f_n \rightarrow f$ on $A$: $\forall \epsilon <0,\ \exists B\subset A:\ \lambda(A\backslash B)< \epsilon$ and $f_n\rightarrow f$ uniformly on $B$.

Because of this uniform convergence I can say that $|f_n -f| < \delta$ for all $\delta>0$ as from some $N$ large enough. Since $A= A\backslash B \cup B$ I have:

$\int_A |f_n -f|dx = \int_{A\backslash B}|f_n -f|dx + \int_B |f_n -f|dx \leq \delta \cdot ( \lambda(B) + \lambda(A\backslash B)) = \delta \cdot \lambda(A) \rightarrow 0$ and hence $\lim_{n\rightarrow \infty} \int_A f_n(x) dx = \int_A f(x) dx$.


The result as stated is false. Let $A=[0,1]$ and $$ f_n(x)=\begin{cases} n &\text{if }0\le x\le1/n,\\ 0 & \text{if }1/n< x\le1. \end{cases}$$ Then $\int_Af_n=1$ for all $n$, but $\lim_{n\to\infty}f_n(x)=0$ for $0<x\le1$.

You certainly have to use the equiintegrability condition.

  • $\begingroup$ Ok, do you have a hint as to how to start the proof? I'm a bit lost at the moment $\endgroup$ – Longeyes Nov 27 '13 at 18:39
  • $\begingroup$ What other convergence theorems do you know? Try to learn from their proofs. $\endgroup$ – Julián Aguirre Nov 27 '13 at 18:52

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.