4
$\begingroup$

Any rearrangement of an absolutely convergent sequence $(a_n)$ is another absolutely convergent sequence with the same limit. Let $(a_{\sigma(n)})$ be the rearranged sequence under the bijection of indices $\sigma$.

My proof attempt. Let $T_N = \sum_{n=1}^N |a_{\sigma(n)}|$ be the $N$th partial sum of the rearranged absolute series. Similarly let $S_N$ be the partial sum of the original absolute series. I know I want to try to prove the absolute sequences Cauchy. Still working on it.

$\endgroup$
  • $\begingroup$ So, not much of an attempt, then? $\endgroup$ – Andrés E. Caicedo Oct 5 '13 at 23:06
  • $\begingroup$ Hold on, I just thought of something $\endgroup$ – BananaCats Category Theory App Oct 5 '13 at 23:07
  • $\begingroup$ Here is the key hint: Given $n$, show that if $N$ is large enough, then all surviving terms in $T_N-S_N$ have index larger than $n$. $\endgroup$ – Andrés E. Caicedo Oct 5 '13 at 23:07
  • $\begingroup$ In think your first two "sequence" in the body of the question must be changed to "series". $\endgroup$ – DonAntonio Oct 5 '13 at 23:15
  • $\begingroup$ @DonAntonio The OP should rather say "absolutely summable". $\endgroup$ – Pedro Tamaroff Oct 5 '13 at 23:16
7
$\begingroup$

You needn't make the $N$s match. You know you can take $N$ large enough so that $$\sum_{k\geqslant N}|a_k|<\varepsilon$$

Now pick $M$ large enough so that all the terms in $T_M$ are some of the terms in $S_N$. Then how large is $$\left|\sum_{k\geqslant 0}a_k-T_M\right|?$$

$\endgroup$
  • $\begingroup$ I don't know the first part. Will have to prove it. $\endgroup$ – BananaCats Category Theory App Oct 6 '13 at 1:42
  • $\begingroup$ @EnjoysMath Check the definition of "absolutely summable" again! $\endgroup$ – Pedro Tamaroff Oct 6 '13 at 1:44

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.