MathGroup Archive 2002

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

Search the Archive

Nintegration - numerical integration converges too slow

  • To: mathgroup at smc.vnet.net
  • Subject: [mg38127] Nintegration - numerical integration converges too slow
  • From: "Fitchard, Ann (PSAF)" <PSAF at chevrontexaco.com>
  • Date: Tue, 3 Dec 2002 04:33:57 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

I am a new Mathematica user with a numerical integration problem.  My 2D
Nintegration gives the warning "slwcon" (numerical integration converges
too slow).  I am integrating over a thin strip through a 2D
Interpolating Function.  The Interpolating function is zero for most
values.  All stips pass through some non-zero region.  When the strip is
mostly in the non-zero region there is no problem.  Howvever, when the
strip passes through only a small part of the non-zero region, the
warning occurs and the integration is slow.  The integration parameters
I have so far found to work best are: WorkingPrecision -> 9,
AccuracyGoal ->Infinity, Precision Goal -> 2, Method -> Gausskronron.
If I increase the PrecisionGoal both the number of warnings and the time
increase.

Can you suggest a better set of parameters.  Thank you in advance.

Ann Fitchard


  • Prev by Date: Re: copying cells in Mathematica 4.2
  • Next by Date: RE: solving non-algebraic exponential equations
  • Previous by thread: Re: Re: copying cells in Mathematica 4.2
  • Next by thread: Re: Nintegration - numerical integration converges too slow