Skip to content

'solve' algorithm

2 messages · W. Beldman, Brian Ripley

#
I'm looking for the algorithm of the  solve  function (base package). As I 
understand the help text neither Linpack nor Lapack's ZESV is used (in the 
simple example below) but what algorithm is? Something like Gauss Jordan 
Elimination should be applicable to solve a linear system

In particular I want to be able to translate it into i.e. Delphi, but in the 
example below I'm not convinced that I understand what's happening. vD seems 
rather small...


TIA,

W. Beldman



 vD <- matrix(0.305625, 1, 1)
[,1]
[1,] 0.305625
[,1]     [,2]     [,3]
[1,] 0.348125 0.296875 0.305625
[,1]      [,2] [,3]
[1,] 1.139059 0.9713701    1
#
On Wed, 14 Jan 2004, W. Beldman wrote:

            
Well, your example is not complex.  By default (LINPACK=FALSE), DGESV is 
used: just read the code.