Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2001
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 2001

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Very simple limit, why doesn't it evaluate?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg31939] Re: Very simple limit, why doesn't it evaluate?
  • From: "Mirek Gruszkiewicz" <gruszkiewicz at ornl.gov>
  • Date: Thu, 13 Dec 2001 01:08:44 -0500 (EST)
  • Organization: Oak Ridge National Lab, Oak Ridge, TN
  • References: <9v77po$s5v$1@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

"Andrzej Kozlowski" <andrzej at bekkoame.ne.jp> wrote in message
news:9v77po$s5v$1 at smc.vnet.net...
> Computing limits is done by means of a number of ad hoc tricks rather
> than a universal algorithm, so there will always be cases that
> Mathematica can't manage but can be done "by hand". In general when
> Mathematica can't find a limit it is a good idea to load in the
> Calculus`Limit` package and try again. In this case this case we get:
>
> << Calculus`Limit`
>
> In[2]:=
> Limit[4^(3 + n)/(-5 + 3^(4 + n)), n -> Infinity]
>
> Out[2]=
> Infinity
>
> This is with Mathematica 4.1 so I am not sure what happens with 3.0 but
> I suspect the same thing.
>

Unfortunately calculating limits is not easy and currently I would not
bother using Mathematica for this purpose, since the results would have to
be verified by hand (brain?) anyway. It might fail to solve a trivial
example, or even give a wrong answer.

For example, the Calculus `Limit` package when asked for

Limit[(3^n)/(3^(n + 1) +1), n -> Infinity]

says 1, while the correct answer is 1/3;

Limit[(3^n)/(3^(n + 2) +1), n -> Infinity]
is given as 1/3 (should be 1/9).

The built-in function does give the correct answers to these, but then it
fails on other equally simple cases.

Mathematica Limit certainly is no Augustin Cauchy and will never be, but it
seems it easily could/should be a little smarter than it is now. It needs
some work.

MG








  • Prev by Date: Preserve subscripts like integers using N
  • Next by Date: Re: Re: question
  • Previous by thread: Re: Very simple limit, why doesn't it evaluate?
  • Next by thread: Re: Split-Step Fourier Method