MathGroup Archive 2006

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

Search the Archive

Re: simplifying inequalities

  • To: mathgroup at smc.vnet.net
  • Subject: [mg72280] Re: [mg72272] simplifying inequalities
  • From: Andrzej Kozlowski <akoz at mimuw.edu.pl>
  • Date: Mon, 18 Dec 2006 06:55:42 -0500 (EST)
  • References: <200612171120.GAA07329@smc.vnet.net>

On 17 Dec 2006, at 20:20, Chris Chiasson wrote:

> Why isn't the result of this command,
>
> FullSimplify[
>   beamLength>5 x&&2 beamLength>5 x&&3 beamLength>5 x&&
>     4 beamLength>5 x,{beamLength>x>0}]
>
> , equal to beamLength>5x?
>
> Mathematica just returns all the inequalities as input. Can something
> be done about this?
>
> --  
> http://chris.chiasson.name/
>

It seems that FullSimplify does not use CylindricalDecomposition in  
such cases:


CylindricalDecomposition[beamLength > 5*x && 2*beamLength > 5*x &&  
3*beamLength > 5*x && 4*beamLength > 5*x && beamLength > 0 && x > 0,
   {x, beamLength}]


x > 0 && beamLength > 5*x


Refine[%,x>0]

beamLength>5 x

CylindricalDecomposition is a function of very high complexity so  
FullSimplify and Simplify use it sparingly. I think Reduce always  
uses CylindricalDecompostion when inequalities are involved so it  
will also work in this case.

Andrzej Kozlowski


  • Prev by Date: Re: Combinatorica parameters
  • Next by Date: Re: simplifying inequalities
  • Previous by thread: simplifying inequalities
  • Next by thread: Re: simplifying inequalities