1
$\begingroup$

I'm new to functional analysis and I'm studying the Banach spaces.

I have some examples where to prove that a space is not Banach/complete, the procedure is to take a Cauchy sequence in the space and showing that it does not converge to an element of the space.

While it seems quite easy to prove that a Cauchy sequence does not converge to an element of the space, to me it seems harder to prove that a chosen sequence is actually a Cauchy sequence. I know the definition of Cauchy seq but I don't understand very well how to use it.

Say we have a normed space $(X,||\cdot||_{L^p})$, we take a sequence $\{f_k\}$ in $X$ and we show that it is bounded in the $L^p$ norm by some integrable function. Moreover we show that the sequence converges pointwise to a function $f$. We have all the hypotesis to use the dominated convergence theorem, which says that $I=\int_X|f_k-f|$ goes to $0$ as $k$ goes to infinity.

The $L^p$ spaces contain all the measurable functions whose $L^p$ norm is finite, and the $L^p$ norm of $f_k-f$ is defined as $J=(\int_X|f_k-f|^p)^{1/p}$. My guess is that since $I\rightarrow0$, then $J\rightarrow0$ too as $k\rightarrow+\infty$.

So, if the guess it true, does this mean that our sequence is Cauchy, i.e. $||f_n-f_m||_{L^p}<\epsilon$ $\forall\epsilon>0,\forall m,n>N\in\mathbb{N}$ ?

$\endgroup$
1
$\begingroup$

I am not sure if I understand the questions completely, but here are some relevant facts: if $f_n \to f$ almost everywhere, $|f_n| \leq g$ and $g \in L^{p}$ then $f \in L^{p}$ and $\int |f_n-f|^{p} \to 0$. Any convergent sequence is always Cauchy, so $\{f_n\}$ is Cauchy in $L^{p}$. [Remark: we have to assume that $g \in L^{p}$. If we just assume that $g$ is integrable then $\{f_n\}$ need not be Cauchy].

$\endgroup$
  • $\begingroup$ Thanks for the reply, is that a theorem or just facts? Because what you stated (if $f_n\rightarrow f, |f_n|\le g\in L^p$ then $f\in L^p, \int|f_n-f|^p\rightarrow0$) seems a slightly modified version of the dominated convergence theorem. $\endgroup$ – sound wave Oct 26 '18 at 13:33
  • $\begingroup$ LDCT hypotesis: $f_n:X\rightarrow\mathbb{R}$ sequence of measurable functions, $f_n\rightarrow f$ for every element in $X$, $|f_n|<g$ where $g:X\rightarrow [0,+\infty]$ is a summable function (i.e. $\int |g|<\infty$, i.e. $g\in L^1)$ $\endgroup$ – sound wave Oct 26 '18 at 13:44
  • $\begingroup$ Ah I've just read that a corollary of the LDCT is the dominated convergence in $L^p$-spaces, which seems exactly what you stated link $\endgroup$ – sound wave Oct 26 '18 at 13:48

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.