-1

I was wondering when is the following statement can be true?

Let $\{x_n\} ,\{y_n\} \subset \mathbb R $ are sequences. then $$ \limsup_n (x_n+y_n) = \limsup_n x_n +\limsup y_n $$

It is clear that it is true if the both of these sequences are convergent.

Also I can see that if $\{y_n \}=$ constant.then the statement above is true.

Also, Let $$ x_n = \left\{ \begin{array}{lr} 1 & \mbox{if } n \text{ is even } \\ 2 & \mbox{if } n \text{ is odd } \end{array} \right.$$ $$ y_n = \left\{ \begin{array}{lr} 2 & \mbox{if } n \text{ is even } \\ 1 & \mbox{if } n \text{ is odd } \end{array} \right.$$

Then, $$ \limsup_n (x_n+y_n) =3 \neq 4= \limsup_n x_n +\limsup y_n$$

So the statement could be false if the both of the sequences are nonconvergent. Also we will have a problem if one of them has $+ \infty $ as a limit and the other one has $- \infty $ as a limit so it isn't true for this case too.

So, only one case left to check, that is if we know that one of these sequence has a limit. And I couldn't find the answer for this case.

I was trying like this :

Since we know that $ \sup(A+B) =\sup A +\sup B $ and $\inf(A+B) =\inf A +\inf B $, then

$$ \limsup_n (x_n+y_n) = \inf_{n \in \mathbb N} \sup_{k \ge n} (x_n +y_n)= \inf_{n \in \mathbb N} \big( \sup_{k \ge n} (x_n) + \sup_{k \ge n} (y_n) \big)=\inf_{n \in \mathbb N} \sup_{k \ge n} (x_n) + \inf_{n \in \mathbb N} \sup_{k \ge n} (y_n) = \limsup_n x_n +\limsup y_n $$
so the statement is true. But here I didn't use the convergence of $\{x_n\}$ at all so it has to be true for any sequences. Since I have already given a counter example then I am sure that my proof is wrong, but I can't find the mistake in this proof. Also I still don't know if the statement is true if one of this sequence is convergent or not.

Any help?

Edit: From the comment I can see the mistake in my proof. So, I have one question now is this statement true when one of the sequences is convergent?

A12345
  • 719
  • 2
    The problem is that you are using a property of supremum/infimum of sum of sets with sum of sequences, in the case of sequences this is not true, since you are not summing all elements of ${x_n}$ with all elements of ${y_n}$ but only the ones with same index. An example would be the a $x_n$ that alternates with $1$ and $-1$ and $y_n = - x_n$ – Raul Fernandes Horta Mar 29 '25 at 19:57
  • @RaulFernandesHorta I got it. Thank you very much. – A12345 Mar 29 '25 at 20:02
  • Generally holds inequality. For equality, for example, is true $\exists\lim\limits_{n\to\infty}x_n\ &\ \forall(y_n){n\in \mathbb{N}} \Rightarrow \varlimsup\limits{n\to\infty}(x_n + y_n ) = \lim\limits_{n\to\infty}x_n + \varlimsup\limits_{n\to\infty}y_n$ – zkutch Mar 29 '25 at 20:47
  • @zkutch how can I prove that? I can't see it – A12345 Mar 29 '25 at 21:03
  • 2
    See here, here and here. Does this answer your questions? – Dean Miller Mar 29 '25 at 22:34
  • 1
    Answer is on second link brought by Dean Miller, but there is no such ones which I prefer: try with subsequences. – zkutch Mar 30 '25 at 01:17
  • 1
    @DeanMiller yes, it does. Thank you very much. – A12345 Mar 30 '25 at 06:58
  • Here is a better match for this post. – Dean Miller Mar 30 '25 at 13:49

1 Answers1

1

I decided to give here a version of the proof for the sentence from the comment using subsequences, since this technique seems more transparent and is not used on brought links above - I hope it will be useful to someone.

So, let's prove $$\exists\lim\limits_{n\to\infty}x_n\ \&\ \forall(y_n)_{n\in \mathbb{N}} \Rightarrow \varlimsup\limits_{n\to\infty}(x_n + y_n ) = \lim\limits_{n\to\infty}x_n + \varlimsup\limits_{n\to\infty}y_n$$ So, we have one limit point for $(x_n)_{n\in \mathbb{N}}$, let's denote it $x=\lim\limits_{n\to\infty}x_n$ and let $\{y^*\}$ be the set of limit points for $(y_n)_{n\in \mathbb{N}}$ i.e. for each $y^*$ exists $(y_{n_k})_{k\in \mathbb{N}}$ converged to it. Sequence $x_n + y_n$, therefore, have limit points $\{y^*+x\}$. For $\varlimsup\limits_{n\to\infty}y_n$ we have some $(y_{n_k})_{k\in \mathbb{N}}$ converged to it, so, $x_{n_k}+y_{n_k}$ converge to $\lim\limits_{n\to\infty}x_n + \varlimsup\limits_{n\to\infty}y_n$. On other hand this subsequence gives $\varlimsup\limits_{n\to\infty}(x_n + y_n )$.

zkutch
  • 14,313
  • This method of proof is already addressed here. Perhaps rather post your answer there (since your explanation has more detail than the others in the linked post) and close this one as a duplicate? – Dean Miller Mar 30 '25 at 13:48
  • @Dean Miller, at the moment when I wrote my answer you and amWhy had not yet provided the last link. Of course, it is not a problem for me to copy this answer there, but the questioner has already accepted it and I do not know how correct it would be from me in relation to him. On the other hand, correct me if I am wrong, isn't in the given link in the accepted answer only proven inequality and not equality? – zkutch Mar 30 '25 at 23:00
  • The reason why I made the suggestion was that there are many other posts about this same question (I found at least four with a bit more searching) and that this question was likely to be closed as a duplicate (which it was). The other post does talk about one direction in particular, but it is not unreasonable to post a proof of the other direction as well. Either way, I do not have a problem with your answer staying here as it is. – Dean Miller Mar 31 '25 at 08:18