MathGroup Archive 1999

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

Search the Archive

Re: Solution of this equation

  • To: mathgroup at smc.vnet.net
  • Subject: [mg20728] Re: Solution of this equation
  • From: Daniel Lichtblau <danl at wolfram.com>
  • Date: Wed, 10 Nov 1999 00:17:43 -0500
  • Organization: Wolfram Research, Inc.
  • References: <7vrc3p$2nd@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

Dave Richardson wrote:
> 
> Can anyone offer insight here?
> 
> This Mathematica expression gives 3 solutions to the equation.
> 
> Solve[Pr == (8*Tr)/(3*vr - 1) - 3/vr^2, vr]
> 
> The problem is that there are 3 Real solutions, and Mathematica is giving
> solutions with (granted a small) imaginary component.
> 
> And hitting it with  a full simplify is just not a good idea...
> 
> Any help?
> 
> Thanks,
> 
> Dave!
> 
> --
> Dave Richardson
> University of Maryland
> Department of Mechanical Engineering
> Center for Environmental Energy Engineering
> (301) 405-8726
> dhr at glue.umd.edu


It may be useful to express the result in terms of Root[...] objects
rather than radicals, as the latter of necessity must be expressed in
terms of Sqrt[-1] for generic cases. To get Root solutions one uses
SetOptions.

In[21]:= SetOptions[Roots, Cubics->False];      

In[22]:= soln = Solve[Pr == (8*Tr)/(3*vr - 1) - 3/vr^2, vr]
                                       2          2          3
Out[22]= {{vr -> Root[-3 + 9 #1 - Pr #1  - 8 Tr #1  + 3 Pr #1  & , 1]}, 
                                  2          2          3
>    {vr -> Root[-3 + 9 #1 - Pr #1  - 8 Tr #1  + 3 Pr #1  & , 2]}, 
                                  2          2          3
>    {vr -> Root[-3 + 9 #1 - Pr #1  - 8 Tr #1  + 3 Pr #1  & , 3]}}

An advantage is that now you get solutions that are known to be
explicitly real when appropriate values for Pr and Tr are used.

In[28]:= vals = vr /. (soln /. {Tr->-8/3, Pr->5/4})
Out[28]= {Root[-36 + 108 #1 + 241 #1  + 45 #1  & , 1], 
                               2        3
>    Root[-36 + 108 #1 + 241 #1  + 45 #1  & , 2], 
                               2        3
>    Root[-36 + 108 #1 + 241 #1  + 45 #1  & , 3]}

In[31]:= Map[Im, vals]
Out[31]= {0, 0, 0}

In[32]:= N[vals]
Out[32]= {-4.82362, -0.752372, 0.220437}

It is also typically faster and more accurate to compute numerical
values for solutions expressed as algebraic numbers rather than in terms
of radicals.


Daniel Lichtblau
Wolfram Research


  • Prev by Date: Re: Re: hexagon tiled torus
  • Next by Date: Re: QR vs. Gram Schmidt
  • Previous by thread: Re: Solution of this equation
  • Next by thread: Mathematica link for LabVIEW