MathGroup Archive 2006

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

Search the Archive

Re: How to get a positive solution from Solve Command

  • To: mathgroup at smc.vnet.net
  • Subject: [mg67825] Re: How to get a positive solution from Solve Command
  • From: "David W. Cantrell" <DWCantrell at sigmaxi.net>
  • Date: Sun, 9 Jul 2006 04:50:52 -0400 (EDT)
  • References: <e8nte8$k9j$1@smc.vnet.net>
  • Sender: owner-wri-mathgroup at wolfram.com

abdou.oumaima at hotmail.com wrote:
> By now, I fix the problem as that:
> f[&#958;_] := Normal[Series[1 - c  ArcTanh[&#958;]/&#958;, {&#958;, 0,
> 2}]] Solve[f[&#958;] == 0, &#958;];
> &#958; = &#958; /. % [[2]];
>
> What do you think?

Given some c in the interval [0,1], you're wanting to solve

1 - c ArcTanh[x]/x == 0 for x.

(Is there some particular application in which that equation arises?)

Unless c happens to be close to 1, the basic method you proposed will give
highly inaccurate answers.

Your equation is equivalent to ArcTanh[x]/x == 1/c, and so if you could
somehow invert the function ArcTanh[x]/x, you could get a neat solution.
As you probably realized, that function cannot be inverted precisely in
closed form. Nonetheless, its inverse can be approximated simply by

Sqrt[1 - Exp[(1 - y)(2 + 1/y)]]

Replacing y with 1/c then gives an approximate solution of your equation
for x:

Sqrt[1 - Exp[(1 - 1/c)(2 + c)]]

For that approximation, |error| < 0.0009 .

David W. Cantrell


  • Prev by Date: help with lists operations
  • Next by Date: Re: Non Linear fit problem
  • Previous by thread: RE:How to get a positive solution from Solve Command
  • Next by thread: RE:How to get a positive solution from Solve Command