Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #10407 > unrolled thread

Splitting sums in mathematica

Started byjimbo <james.a.gordon1@googlemail.com>
First post2012-06-09 07:09 +0000
Last post2012-06-14 09:30 +0000
Articles 6 — 3 participants

Back to article view | Back to comp.soft-sys.math.mathematica


Contents

  Splitting sums in mathematica jimbo <james.a.gordon1@googlemail.com> - 2012-06-09 07:09 +0000
    Re: Splitting sums in mathematica Ray Koopman <koopman@sfu.ca> - 2012-06-10 06:14 +0000
      Re: Splitting sums in mathematica Ray Koopman <koopman@sfu.ca> - 2012-06-11 04:04 +0000
        Re: Splitting sums in mathematica Ray Koopman <koopman@sfu.ca> - 2012-06-13 08:54 +0000
    Re: Splitting sums in mathematica "james.a.gordon1" <james.a.gordon1@googlemail.com> - 2012-06-13 09:00 +0000
      Re: Splitting sums in mathematica Ray Koopman <koopman@sfu.ca> - 2012-06-14 09:30 +0000

#10407 — Splitting sums in mathematica

Fromjimbo <james.a.gordon1@googlemail.com>
Date2012-06-09 07:09 +0000
SubjectSplitting sums in mathematica
Message-ID<jqusrg$rnm$1@smc.vnet.net>
Hello,

Does anyone know of a way to algebraically split sums (if that is the
correct name for it) in mathematica? I need to do the following (and
higher-order versions of the same)...

Sum_over_p [  Sum_over_q (  a_p * a_q  )   ] = Sum_over_p_equal_q
(a_p^2) + Sum_over_p_not_equal_q (a_p*a_q)

...in other words, turn a nested sum running over two indices into two
sum, one for when the indices are equal, and one for when they are
different?

I then need to make some substitutions once there are no nested sums
and simplify my answer. I have an equation with many hundreds of terms
in it, so doing it by hand is not an option. I can send a typeset
example if anyone needs clarification

I have version 5.2 for students

Thanks in advance

James

[toc] | [next] | [standalone]


#10421

FromRay Koopman <koopman@sfu.ca>
Date2012-06-10 06:14 +0000
Message-ID<jr1dvt$73t$1@smc.vnet.net>
In reply to#10407
On Jun 9, 12:09 am, jimbo <james.a.gord...@googlemail.com> wrote:
> Hello,
>
> Does anyone know of a way to algebraically split sums (if that is the
> correct name for it) in mathematica? I need to do the following (and
> higher-order versions of the same)...
>
> Sum_over_p [  Sum_over_q (  a_p * a_q  )   ] = Sum_over_p_equal_q
> (a_p^2) + Sum_over_p_not_equal_q (a_p*a_q)
>
> ...in other words, turn a nested sum running over two indices into two
> sum, one for when the indices are equal, and one for when they are
> different?
>
> I then need to make some substitutions once there are no nested sums
> and simplify my answer. I have an equation with many hundreds of terms
> in it, so doing it by hand is not an option. I can send a typeset
> example if anyone needs clarification
>
> I have version 5.2 for students
>
> Thanks in advance
>
> James

Let a = {a1,a2,...}. Then a.a gives the sum of the terms with p == q,
and 2*Dot@@Transpose@Subsets[a,{2}] gives the sum of the terms with
p != q.

[toc] | [prev] | [next] | [standalone]


#10427

FromRay Koopman <koopman@sfu.ca>
Date2012-06-11 04:04 +0000
Message-ID<jr3qni$gr3$1@smc.vnet.net>
In reply to#10421
On Jun 9, 11:14 pm, Ray Koopman <koop...@sfu.ca> wrote:
> On Jun 9, 12:09 am, jimbo <james.a.gord...@googlemail.com> wrote:
>> Hello,
>>
>> Does anyone know of a way to algebraically split sums (if that is the
>> correct name for it) in mathematica? I need to do the following (and
>> higher-order versions of the same)...
>>
>> Sum_over_p [  Sum_over_q (  a_p * a_q  )   ] = Sum_over_p_equal_q
>> (a_p^2) + Sum_over_p_not_equal_q (a_p*a_q)
>>
>> ...in other words, turn a nested sum running over two indices into two
>> sum, one for when the indices are equal, and one for when they are
>> different?
>>
>> I then need to make some substitutions once there are no nested sums
>> and simplify my answer. I have an equation with many hundreds of terms
>> in it, so doing it by hand is not an option. I can send a typeset
>> example if anyone needs clarification
>>
>> I have version 5.2 for students
>>
>> Thanks in advance
>>
>> James
>
> Let a = {a1,a2,...}. Then a.a gives the sum of the terms with p == q,
> and 2*Dot@@Transpose@Subsets[a,{2}] gives the sum of the terms with
> p != q.

If you have to do that many times then you might want to define
{p,q} = Transpose@Subsets[Range@Length@a,{2}]  once, and then
use  2*a[[p]].a[[q]]  to sum the terms whose subscripts differ.

[toc] | [prev] | [next] | [standalone]


#10445

FromRay Koopman <koopman@sfu.ca>
Date2012-06-13 08:54 +0000
Message-ID<jr9kfc$3i3$1@smc.vnet.net>
In reply to#10427
On Jun 10, 9:04 pm, Ray Koopman <koop...@sfu.ca> wrote:
> On Jun 9, 11:14 pm, Ray Koopman <koop...@sfu.ca> wrote:
>> On Jun 9, 12:09 am, jimbo <james.a.gord...@googlemail.com> wrote:
>>> Hello,
>>>
>>> Does anyone know of a way to algebraically split sums (if that is the
>>> correct name for it) in mathematica? I need to do the following (and
>>> higher-order versions of the same)...
>>>
>>> Sum_over_p [  Sum_over_q (  a_p * a_q  )   ] = Sum_over_p_equal_q
>>> (a_p^2) + Sum_over_p_not_equal_q (a_p*a_q)
>>>
>>> ...in other words, turn a nested sum running over two indices into two
>>> sum, one for when the indices are equal, and one for when they are
>>> different?
>>>
>>> I then need to make some substitutions once there are no nested sums
>>> and simplify my answer. I have an equation with many hundreds of terms
>>> in it, so doing it by hand is not an option. I can send a typeset
>>> example if anyone needs clarification
>>>
>>> I have version 5.2 for students
>>>
>>> Thanks in advance
>>>
>>> James
>>
>> Let a = {a1,a2,...}. Then a.a gives the sum of the terms with p == q,
>> and 2*Dot@@Transpose@Subsets[a,{2}] gives the sum of the terms with
>> p != q.
>
> If you have to do that many times then you might want to define
> {p,q} = Transpose@Subsets[Range@Length@a,{2}]  once, and then
> use  2*a[[p]].a[[q]]  to sum the terms whose subscripts differ.

The sum of the terms whose subscripts differ is  Tr@a^2 - a.a,
and it may be faster to let Mathematica simplify that.

[toc] | [prev] | [next] | [standalone]


#10449

From"james.a.gordon1" <james.a.gordon1@googlemail.com>
Date2012-06-13 09:00 +0000
Message-ID<jr9kqr$3m8$1@smc.vnet.net>
In reply to#10407
Hi Koopman,

Thanks for the quick reply, I believe your suggestions solve some of my problem. To further clarify my question, I have an equation that looks like this

F = (X - Y1 - Y2 - Y3 + Z)^2

X = Sum_p Sum_q Sum_r [a_p*a_q*a_r]

Yi = Sum_p Sum_q!=p b_p*b_q, where different X_i have different terms in the sums

Z = Sum_p Sum_q!=p Sum_r!=q!=p

In other words I end up with a large number of terms, each of which can have as many as 6 nested sums, all of which need to be expanded resulting in hundreds of terms. I want mathematica to algebraically expand these nested sums for me rather than me having to type up the expanded form myself. I believe with your solution I would still need to manually keep track of the indices for each expansion. 

Once all the nested sums are expanded, I will make some substitutions of the a, a^2, a^3...a^6 terms, then simplify. At no point to I need to calculate this with real values, I just need an algebraic expression.

Thanks again

James


[toc] | [prev] | [next] | [standalone]


#10451

FromRay Koopman <koopman@sfu.ca>
Date2012-06-14 09:30 +0000
Message-ID<jrcav2$dpb$1@smc.vnet.net>
In reply to#10449
On Jun 13, 2:00 am, "james.a.gordon1" <james.a.gord...@googlemail.com>
wrote:
> Hi Koopman,
>
> Thanks for the quick reply, I believe your suggestions solve some
> of my problem. To further clarify my question, I have an equation
> that looks like this
>
> F = (X - Y1 - Y2 - Y3 + Z)^2
>
> X = Sum_p Sum_q Sum_r [a_p*a_q*a_r]
>
> Yi = Sum_p Sum_q!=p b_p*b_q, where different X_i have different
> terms in the sums
>
> Z = Sum_p Sum_q!=p Sum_r!=q!=p
>
> In other words I end up with a large number of terms, each of which
> can have as many as 6 nested sums, all of which need to be expanded
> resulting in hundreds of terms. I want mathematica to algebraically
> expand these nested sums for me rather than me having to type up
> the expanded form myself. I believe with your solution I would still
> need to manually keep track of the indices for each expansion.
>
> Once all the nested sums are expanded, I will make some
> substitutions of the a, a^2, a^3...a^6 terms, then simplify.
> At no point to I need to calculate this with real values,
> I just need an algebraic expression.
>
> Thanks again
>
> James

I don't understand what you're trying to do -- how the Yi differ
from one another, or what it means to "make some substitutions
of the a, a^2, a^3...a^6 terms", as opposed to substituting for
individual terms. Give us a small example, using Mathematica syntax.

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web