Let $(X , \mathcal{A} , \mu)$ a finite measurable space and $ \mathcal{F}$ the of all $\mathcal{A}$-measurable functions $f : X \rightarrow \mathbb{R} $ .For $ f , g \in \mathcal{F}$ we define : $$ d(f , g) = \int_X \frac{| f -g|}{1 +|f - g| } d \mu .$$

So we have :

  1. $d (f , g) = 0 \Rightarrow f = g $ a.e.
  2. $d(f,g) = d(g,f)$
  3. $d(f,g) \leq d (f,h) + d(h,g)$

    We want to show that if $f_n \in \mathcal{F}$ , for $n = 1,2 , \dots$ is such that $d(f_m , f_n) \rightarrow 0$ so there is a $f \in \mathcal{F}$ such that $d(f_n , f) \rightarrow 0$.

My try : as $d(f_m , f_n) \rightarrow 0 $ we have that for any given $ k \in \mathbb{N}$ there is $n_0 \in \mathbb{N}$ which if $m ,n > n_0$ $d(f_m , f_n) < 1/k$, so $|f_m - f_n| < 1/k $ a.e. , so there is $R_k \subset X$ that $|f_m(x) - f_n(x)| < 1/k $ for all $x \in R_k$ and $\mu (X \backslash R_k) = 0$. Taking $ R = \bigcap_k R_k$, and as $(f_n(x))$ is as Cauchy sequence in $\mathbb{R}$ there is $\lim f(x) $for all $x \in R$, so we take $f(x ) = \lim f(x)$, for $x \in R$. But I don't know if will work. Thanks in advance!

  • $\begingroup$ $d(f_m,f_n)<1/k$ does not imply $|f_m-f_n|<1/k$ a.e.. $\endgroup$ – Luiz Cordeiro Sep 14 '16 at 22:31

Given $\epsilon>0$, consider the set $A = \{|f_m(x)-f_n(x)|>\epsilon\}$. Then \begin{equation} d(f_n,f_m)>\frac{\epsilon}{1+\epsilon}\mu(A), \end{equation} which implies that $\mu(A)\to 0$ as $m,n\to 0$, i.e. $f_n$ is a Cauchy sequence in measure. Then there exists an $f$ such that $f_n\to f$ in measure.

Cauchy in measure implies convergent in measure.

Again we have for $B=\{|f_m(x)-f(x)|>\epsilon\}$ \begin{equation} d(f_n,f)\leq \frac{\epsilon}{1+\epsilon}\mu(X) + 1 \mu(B), \end{equation} as, $\mu(X)$ is finite, $\epsilon$ is arbitrary and $\mu(B)\to 0$, we are done.


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.