Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.soft-sys.math.maple > #335
| From | Peter Pein <petsie@dordos.net> |
|---|---|
| Newsgroups | comp.soft-sys.math.maple |
| Subject | Re: Series Summation Question |
| Date | 2012-01-27 07:55 +0100 |
| Organization | 1&1 Internet AG |
| Message-ID | <jfthpq$18d$1@online.de> (permalink) |
| References | <1fe7f908-df47-4b54-949f-630a50b0551e@j14g2000vba.googlegroups.com> <87vcnywi1u.fsf@san.rr.com> <39d1ff7b-301a-48ee-99a6-3e4d0df84b7a@o13g2000vbf.googlegroups.com> <jftgbn$bt$1@online.de> |
Am 27.01.2012 07:31, schrieb Peter Pein:
> Am 27.01.2012 01:45, schrieb clashton:
>> On Jan 26, 11:21 am, Joe Riel<j...@san.rr.com> wrote:
>>> clashton<clash...@gmail.com> writes:
>>>> I am a little curious about whether Maple can evaluate a certain
>>>> infinite sum that fails with Mathematica (I do not have Maple so
>>>> cannot test it myself).
>>>
>>>> Here is the sum in Mathematica:
>>>> Sum[q^(-6 + 4*n)/(1 - q^(-5 + 4*n)), {n, 0, Infinity}] /. {q -> 0.1}
>>>> This leads to and error,
>>>> but
>>>> (Sum[q^(-6 + 4*n)/(1 - q^(-5 + 4*n)), {n, 0, 1}] /. {q -> 0.1)
>>>> +
>>>> (Sum[q^(-6 + 4*n)/(1 - q^(-5 + 4*n)), {n, 2, Infinity}] /. {q -> 0.1})
>>>> gives the correct answer, which I find a little strange.
>>>
>>>> How does Maple perform on the original sum?
>>>
>>> Summing numerically (and guessing at the meaning of the Mathematica):
>>>
>>> (**) S := Sum(q^(-6 + 4*n)/(1 - q^(-5 + 4*n)), n=0..infinity);
>>> infinity
>>> ----- (-6 + 4 n)
>>> \ q
>>> S := ) ---------------
>>> / (-5 + 4 n)
>>> ----- 1 - q
>>> n = 0
>>>
>>> (**) evalf(eval(S,q=1/10));
>>> -21.10120010
>>>
>>> --
>>> Joe Riel
>>
>> Maple wins this round.
>
> Well, if you are interested in the approximation of the sum's value for
> q=1/10, you should have used NSum in Mathematica:
>
> NSum[q^(-6 + 4*n)/(1 - q^(-5 + 4*n)) /. q->1/10, {n, 0, Infinity},
> WorkingPrecision->20]
>
> -21.101200102001001100
>
> or - using your workaround - get
>
> 6/(5*(-1 + q)) + (1 + 2*q + 3*q^2 + 4*q^3)/(5*(1 + q + q^2 + q^3 + q^4))
> + (-2*Log[q^4] + Log[1 - q^4] + QPolyGamma[0, 2 - Log[q^5]/Log[q^4],
> q^4])/(q*Log[q^4])
>
> as (hopefully) exact value.
>
> Cheers, Peter
Sorry for posting too fast. One gets the result in a more simple form by
doing sth. more complicated:
In[1]:= f[q_] = q^(-6 + 4*n)/(1 - q^(-5 + 4*n));
assum = SumConvergence[f[q], n]
s[q_] = Together[Subtract @@ (Limit[Sum[f[q], n], n -> #1, Assumptions
-> assum] & ) /@ {Infinity, 0}]
N[s[1/10]]
Out[2]= q != 0 && Abs[q]^4 < 1
Out[3]= (Log[1 - q^4] + QPolyGamma[0, -(Log[q^5]/Log[q^4]),
q^4])/(q*Log[q^4])
Out[4]= -21.1012
Back to comp.soft-sys.math.maple | Previous | Next — Previous in thread | Next in thread | Find similar
Series Summation Question clashton <clashton@gmail.com> - 2012-01-25 22:46 -0800
Re: Series Summation Question Axel Vogt <&noreply@axelvogt.de> - 2012-01-26 19:41 +0100
Re: Series Summation Question Joe Riel <joer@san.rr.com> - 2012-01-26 08:21 -0800
Re: Series Summation Question clashton <clashton@gmail.com> - 2012-01-26 16:45 -0800
Re: Series Summation Question Peter Pein <petsie@dordos.net> - 2012-01-27 07:31 +0100
Re: Series Summation Question Peter Pein <petsie@dordos.net> - 2012-01-27 07:55 +0100
Re: Series Summation Question clashton <clashton@gmail.com> - 2012-01-28 07:59 -0800
Re: Series Summation Question Axel Vogt <&noreply@axelvogt.de> - 2012-01-28 20:21 +0100
Re: Series Summation Question Peter Pein <petsie@dordos.net> - 2012-02-02 12:21 +0100
csiph-web