MathGroup Archive 2001

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

Search the Archive

Cauchy problem PDE NDSolve

  • To: mathgroup at smc.vnet.net
  • Subject: [mg31002] Cauchy problem PDE NDSolve
  • From: "Eugene" <evs at po.cwru.edu>
  • Date: Mon, 1 Oct 2001 02:44:27 -0400 (EDT)
  • Organization: CWRU
  • Sender: owner-wri-mathgroup at wolfram.com

I am trying to solve a parabolic PDE of a Fokker-Plank type that describes
time evolution of a Gaussian peak in infinite space with NDSolve (Math 3.0).
Obviously, infinite space is the problem, and NDSolve generates something
weird.  I played with the NDSolve parameters and various transformations,
but it didn't help (though the equation coefficients are good smooth
functions).  Does anybody know a Mathematica program/package for a Cauchy
parabolic (diffusion-type) PDE problem in infinite space (asymptotically
vanishing boundary conditions)?  I hope it should exist because NDSolve in
the present form fails to solve even a textbook problem of the diffusion
evolution of a Gaussian peak if the diffusion coefficient changes slightly
in space (it generates some mesh-dependent crap).

Thanks!
Eugene





  • Prev by Date: help: how do you solve a less-or-equal equation
  • Next by Date: Re: Mathematica 4.0 installation problem
  • Previous by thread: RE: help: how do you solve a less-or-equal equation
  • Next by thread: Re: Mathematica 4.0 installation problem