3
$\begingroup$

How can I find this limit (where $n$ is an integer number) in WM v 12.2: $\lim_{n\to+\infty} \sin^2(\pi\,\sqrt{n^2+n})$?

I mean, it's clear that it equals to one. Versions 9 and below easily produced the correct answer. Now in v12.2 I've a lot of problems finding the same limit. Any idea? Thank you

$\endgroup$
3
  • 1
    $\begingroup$ Well correct me if i'm wrong, but its not clearly zero. It's actually one. First: this is what you can guess numerically. Second, you'll see that Sqrt[n^2+n] is growing asymptotically like n+1/2. See Limit[Sqrt[n^2+n]-n,n->Infinity] giving 1/2. Thus we can pull the limit into the sine squared and get asymptotically Sin[Pi*(n+1/2)]^2 giving one, as seen numerically too. Did I do something wrong? Btw. although it does not give a result here, DiscreteLimit is Mathematicas function for Limits over integers. $\endgroup$ Commented May 7, 2021 at 14:44
  • $\begingroup$ Sorry, sure it's one. Of course. Mea kulpa. But the problem is that DiscreteLimit does not produce this answer. $\endgroup$ Commented May 7, 2021 at 14:54
  • $\begingroup$ DiscreteLimit works finally in 13.2!! $\endgroup$ Commented Dec 27, 2022 at 13:12

2 Answers 2

3
$\begingroup$

This can be done in 12.2 as follows.

DiscreteAsymptotic[Sin[Pi*Sqrt[n^2 + n]]^2, n -> Infinity]

1

$\endgroup$
3
  • $\begingroup$ Thank you, that's it! $\endgroup$ Commented May 7, 2021 at 14:56
  • $\begingroup$ In 13.1 DiscreteLimit[Sin[Pi*Sqrt[n^2 + n]]^2, n -> Infinity] is still unevaluated. $\endgroup$
    – user64494
    Commented Jul 7, 2022 at 15:15
  • $\begingroup$ But it 13.2 it works!!! Strange, but cloud is now much faster... $\endgroup$ Commented Dec 27, 2022 at 13:11
1
$\begingroup$

I suppose it is not equal to one if it is over Reals:

Asymptotics`ClassicLimit[Sin[Pi*Sqrt[n^2 + n]]^2, n -> Infinity]

prints Interval[{0, 1}]. While

Asymptotics`ClassicLimit[Sin[Pi*Sqrt[n^2 + n]]^2, n -> Infinity, 
 Assumptions -> n \[Element] Integers]

does indeed print one. That is what versions 11.2 and below were using.

I suppose that is like one of those examples in DiscreteLimit[] documentation, after all "integers only" is far from the same as "all reals" limit.

(We cannot use modern Limit to solve this, since it does not allow any Assume[] or Assumptions if it is on the variable we are limiting.)

Edit:

In 13.2 it finally works!

DiscreteLimit[Sin[Pi*Sqrt[n^2 + n]]^2, n -> Infinity] 

prints 1.

$\endgroup$
3
  • $\begingroup$ You open an opened door. $\endgroup$
    – user64494
    Commented Jul 7, 2022 at 15:13
  • $\begingroup$ So you do not agree the result over reals is not just 1 number, but dense interval (well, in common sense limit does not exist)? $\endgroup$ Commented Jul 7, 2022 at 16:04
  • $\begingroup$ @user64494 This works now, did you notice? $\endgroup$ Commented Dec 6, 2023 at 23:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.