Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
2002
*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 2002

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

Search the Archive

Re: Question about yet another bug in Sum

  • To: mathgroup at smc.vnet.net
  • Subject: [mg35222] Re: [mg35204] Question about yet another bug in Sum
  • From: Andrzej Kozlowski <andrzej at tuins.ac.jp>
  • Date: Wed, 3 Jul 2002 05:13:20 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

I am not sure what exactly you mean by "the origin of the bug" (writing 
code too late at night?) but I can tell you for sure that it is do do 
with Limit (as usual, one might add). How do I know this? Well:

<< Calculus`Limit`

Now:

In[2]:=
Sum[Exp[(-I)*Pi*n]*(n/(1 + n^2)), {n, 0, Infinity}]

Out[2]=
(1/4)*PolyGamma[0, 1/2 - I/2] + (1/4)*PolyGamma[0, 1/2 + I/2] -
   (1/4)*PolyGamma[0, 1 - I/2] - (1/4)*PolyGamma[0, 1 + I/2]

A nice closed for answer. We can check it agrees numerically with yours:

In[3]:=
N[%]//Chop

Out[3]=
-0.269611

In fact we can check it symbolically:

In[4]:=
Unevaluated[(1/4)*PolyGamma[0,1/2-I/2]+(1/4)*PolyGamma[0,1/2+I/2]-(1/4)*
           PolyGamma[0,1-I/2]-(1/4)*PolyGamma[0,1+I/2]==-1/2*
         HypergeometricPFQ[{2,1+I,1-I},{2-I,2+I},-1]]//FullSimplify

Out[4]=
True


So life is not so bad after all.



Andrzej Kozlowski
Toyama International University
JAPAN
http://platon.c.u-tokyo.ac.jp/andrzej/


On Tuesday, July 2, 2002, at 03:11  PM, Vladimir Bondarenko wrote:

> Hello.
>
> Mathematica is a great environment. I observe that every time before it
> becomes stronger, Mathematica expeciences a collapse ;-)
> Now we at this leg of its majestic trajectory...
>
> Here is another bug in Sum.
>
>             $Version
>
>             "4.2 for Microsoft Windows (February 28, 2002)"
>
>             Sum[Exp[-I Pi n] n /(1 + n^2), {n, 0, Infinity}]
>
> ACTUAL:     Sum::"div": "Sum does not converge."
>
> EXPECTED:   -1/2*HypergeometricPFQ[{2, 1 + I, 1 - I}, {2 - I, 2 + I}, 
> -1]
>
> CHECK-UP:   N[%, 20]//Chop
>
>             -0.26961050270800898180
>
>             NSum[Exp[-I Pi n] n/(1 + n^2), {n, 0, Infinity}, 
> WorkingPrecision->20]//Chop
>
>             -0.269610502708010
>
> This bug is present in 4.1, 4.0, 3.0.  Version 2.2 leaves the sum 
> unevaluated.
>
>
> Can anyone explain the origin of the bug?
>
>
> Best,
>
> Vladimir Bondarenko
>
>
>
>
>



  • Prev by Date: RE: non integer exponents in ContourPlot[]?
  • Next by Date: Re: I NEED HELP WITH THIS
  • Previous by thread: Re: Question about yet another bug in Sum
  • Next by thread: Simplifying the *Individual Coefficients* in Series Expansions?