I encountered this problem while studying a past comprehensive exam.

Let $(X,\mathcal{M}, \mu)$ be a finite measure space. Given measurable $f: X \rightarrow \mathbb{C}$, let $E_f(\lambda) = \mu(\{x: |f(x)| > \lambda\})$ be its distribution function. Suppose that $\{f_n\}$ is a sequence of measurable functions converging to $f$ $\mu$-a.e. such that there exists a lebesgue measurable function $E(\lambda)$ such that $E: (0,\infty) \rightarrow [0,\infty)$ satisfying $E_{f_n}(\lambda) < E(\lambda)$ and $\int_{(0,\infty)}E(\lambda)d\lambda < \infty$ . Prove that $f_n$ converges to $f$ in $L^1$.

I suspect that Egoroff's theorem needs to be used, and that I need to somehow control the integral of $|f-f_n|$ on the exceptional set using $E(\lambda)$ but I can't get much further than that. I'm not even sure for example why $f\in L^1$ even though the conditions on $E_{f_n}$ force $f_n$ to be.


Since $f_n$ converges to $f$ $\mu$-almost everywhere, it suffices to show that $f, f_n\in L^1$ and that $lim_{n \rightarrow \infty}{||f_n||_1} = ||f||_1$. Now $lim_{n \rightarrow \infty}{||f||_1} = lim_{n \rightarrow \infty}{\int_{(0,\infty)}{E_{f_n}(\lambda)~d\lambda}} = \int_{(0,\infty)}{lim_{n\rightarrow \infty}{E_{f_n}(\lambda)}~d\lambda}$ by dominated convergence.

By Egoroff's theorem, $f_n$ converges to $f$ in measure, therefore $E_{f_n}$ converges to $E_{f}$ at every point of continuity. Since $E_{f_n}$ is right continuous, it has at most countably many discontinuities. therefore $lim_{n\rightarrow \infty}{E_{f_n}(\lambda)} = E_f(\lambda)$ almost everywhere.

So $\int_{(0,\infty)}{lim_{n\rightarrow \infty}{E_{f_n}(\lambda)}~d\lambda} = \int_{(0, \infty)}{E_{f}(\lambda)~d\lambda} = ||f||_1$. Since everything is finite, this concludes the proof.


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.