MathGroup Archive 1998

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

Search the Archive

RE: Mantain unevaluated

  • To: mathgroup at smc.vnet.net
  • Subject: [mg14599] RE: [mg14551] Mantain unevaluated
  • From: Daniel CLEMENT <dclement at mail.cpod.fr>
  • Date: Mon, 2 Nov 1998 01:51:07 -0500
  • Sender: owner-wri-mathgroup at wolfram.com

Guillermo,

Have you tried to Hold[...] the calculation that you want to remain
unevaluated?

Daniel CLEMENT

> ----- Original Message -----
> From:	Guillermo de Paz [SMTP:guipaz at dali.eis.uva.es]
> Sent:	Friday, October 30, 1998, 9:07:25
> To:	mathgroup at smc.vnet.net
> Subject:	[mg14551] Mantain unevaluated
> 
> I'm using Mathematica to obtain the differential equations of some
> mechanical systems and to integrate them numerically. When constructing
> the equations I need to invert a matrix. This matrix is not very big
> (lets say 8x8) but the elements al very long trigonometric expressions,
> the inverse takes a long time. By the other side, it is possible that
> for some values of the dependent variables the rank of the matrix is
> not complete, and the size of the problem decreases. For these reasons
> I would like to have unevaluated this inverse until the numerical
> function NDSolve uses it (not to compute symbolically the matrix, but
> compute it numerically inside NDSolve in each integration step).
> I have done some attempts with := and with Compile, but I cant avoid the
> symbolic  computation of the inverse. Could somebody help me?
> 
> Guillermo de Paz
> Universidad de Valladolid.
> e-mail: guipaz at dali.eis.uva.es
> 
> 
> ----- End Of Original Message -----




  • Prev by Date: Re: mg[14566]: Follow-up on Piecewise Continuous Functions
  • Next by Date: Mathematica - Which Version To Buy
  • Previous by thread: Re: Mantain unevaluated
  • Next by thread: Discontinuity of theta functions