MathGroup Archive 2001

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

Search the Archive

LinearSolve, Det and accuracy

  • To: mathgroup at smc.vnet.net
  • Subject: [mg30648] LinearSolve, Det and accuracy
  • From: Eric Barber <sizex at idi.ntnu.no>
  • Date: Sun, 2 Sep 2001 03:58:54 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

I've been using LinearSolve in Mathematica to get a solution for a
Markov probability system where the coefficients range over 6
magnitudes. A colleague is trying to use CalculationCenter to run my
notebook. Unfortunately CalculationCenter doesn't have LinearSolve, so
we've been trying to get a solution using Det. The naive approach using
Det[matrix with reduced column]/Det[matrix] for each column in turn
isn't able to get a solution - well, it bunches the probability mass in
two states. are there alternative approaches in CalculationCenter;
presumable LinearSolve's multiple methods solution techniques are able
to cope with a system where a naive approach doesn't work.

thanks, Eric


  • Prev by Date: Good Looking Dashed Lines?
  • Next by Date: [Fwd: FYI: Mathematica forces re-indexing hard disk]
  • Previous by thread: Good Looking Dashed Lines?
  • Next by thread: Re: LinearSolve, Det and accuracy