MathGroup Archive 2005

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

Search the Archive

Re: Simplification to Partial Fractions

  • To: mathgroup at smc.vnet.net
  • Subject: [mg59717] Re: [mg59676] Simplification to Partial Fractions
  • From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
  • Date: Thu, 18 Aug 2005 00:17:32 -0400 (EDT)
  • References: <001c01c5a383$320181d0$6901a8c0@JonathanPalmer> <4E9678AE-12C7-43A2-B1BE-14EE7DF2AB10@mimuw.edu.pl>
  • Sender: owner-wri-mathgroup at wolfram.com

Actually, it is easy to see  that if the problem is as you stated it  
then there are infinitely many solutions. So you really need to  
strongly restrict the meaning of "partial fraction". For example, it  
is usually required that the degree of the numerator is strictly less  
than that of the denominator.

Acdrzej Kozlowski


On 17 Aug 2005, at 15:05, Andrzej Kozlowski wrote:

> I don't think you will be able to do it since your representation  
> is not unique. Given your u2 you can produce lot's of different  
> representations of the form u1. For exmple, here is another one:
>
>
> (B*(x^2 - 3*y^2)*(x^2 + z^2) + C*(y^2 + z^2)*(x^2 + z^2) +
>     D*(5*x^4 + 2*z^2*x^2 + z^4))/(x^2 + z^2) + A -
>   (4*(D*x^4 + C*y^2*x^2 + D*y^2*x^2 - B*y^4 + C*y^4))/(x^2 + y^2) +
>   (4*C*y^4)/(y^2 + z^2)
>
> It is less nice than yours but has "the same form". In order to be  
> able to reverse the process of expanding your original expression  
> should be in some sense unique (unless you want to find all such  
> representations, which will probably be computationally v. hard)  
> and unless you can specify more usable information this won't be  
> possible.
>
> Andrzej Kozlowski
>
>
>
>
> On 18 Aug 2005, at 01:27, Jon Palmer wrote:
>
>
>> At first glance I can't make PolynomialReduce do what I need.
>>
>> Here is an example problem. Take the expression:
>>
>> u1 = A + (B*(x^2 - y^2)^2)/(x^2 + y^2) +  (C*(y^2 - z^2)^2)/(y^2 +  
>> z^2)
>>  + (D*(-x^2 + z^2)^2)/(x^2 + z^2)
>>
>> Now
>>
>> u2 = Factor[u1]
>>
>>
>>
>> How do you Simplify u2 back to the form of u1?
>>
>> Many thanks
>> Jon Palmer
>>
>>
>>
>>
>>
>>
>>> -----Original Message-----
>>> From: Andrzej Kozlowski [mailto:akoz at mimuw.edu.pl]
To: mathgroup at smc.vnet.net
>>> Sent: Wednesday, August 17, 2005 11:14 AM
>>> To: Jon Palmer
>>> Cc: mathgroup at smc.vnet.net
>>> Subject: [mg59717] Re: [mg59676] Simplification to Partial Fractions
>>>
>>>
>>> On 17 Aug 2005, at 10:00, Jon Palmer wrote:
>>>
>>>
>>>
>>>> I was wondering if someone can help with a Partial Fraction  
>>>> problem.
>>>>
>>>> I have a calculated expression, u, which is a quotient of two
>>>> polynomials in
>>>> three variables x, y & z.
>>>>
>>>>
>>>> u = P(x,y,z)/Q(x,y,z)
>>>>
>>>>
>>>> I know that the quotient, when simplified, is a sum of partial
>>>> fractions of
>>>> the form
>>>>
>>>> u = R(x,y,z) + S(x,y,z)/(x^2 +y^2)  + T(x,y,z)/(y^2 +z^2) + U
>>>> (x,y,z)/(z^2
>>>> +x^2)
>>>>
>>>>
>>>> Is there a way to simplify the expression into the parial fraction
>>>> form?
>>>>
>>>> I have tried various combinations of Simplify, Apart, Collect etc.
>>>> and can't
>>>> find a method that works. Any help would be much appreciated.
>>>>
>>>> Thanks
>>>> Jon Palmer
>>>>
>>>>
>>>>
>>>>
>>>
>>> It should be possible to do this using PolynomialReduce but you  
>>> would
>>> have to post the actual problem before I could tell for sure.
>>>
>>> Andrzej Kozlowski
>>>
>>>
>>
>>
>>
>>
>
>


  • Prev by Date: Re: Simplification to Partial Fraction
  • Next by Date: Re: Mathematica takes wrong derivative
  • Previous by thread: Re: Simplification to Partial Fractions
  • Next by thread: Re: Simplification to Partial Fractions