Re: [Bug-glpk] Incorrect result from glpk-4.39

[Ali Baharev]

Hello,

To begin with, i cannot help you.

It seems to me that your problem has to do something with scaling.

>  A: min|aij| =  1.233e-32  max|aij| =  6.136e-01  ratio =  4.978e+31

If you look at the ratio, it is clear that the difference 6.136e-01 -
1.233e-32 will have zero correct significant digits when computed with
double precision arithmetic.

I have reported a bug recently, which also has to do something with scaling.

http://lists.gnu.org/archive/html/bug-glpk/2009-08/msg00000.html

I hope Andrew can resolve the problems soon.

Ali