Re: [Bug-glpk] numerical instability (cycling?)

[Ali Baharev]

Dear Andrew,

Thanks for the help.

> I do not think that another scaling algorithm would help. The defect
> is in the simplex solver.

Could you please fix that? Would it be a lot of work?

>> How can i figure out what "tiny" is?
>
> It depends on the instance's nature. I could suggest the following
> criteria: a[i,j] is tiny if |a[i,j]| < 1e-8 * max|a[i,*]| assuming
> that max|a[i,*]| is not very huge. For example, in the row:
>
> 1.23 x1 + 2.34 x2 + 3.45e-15 x3 + 4.56 x4 <= 5.67
>
> constraint coefficient '3.45e-15' is tiny and should be replaced by
> exact zero (or skipped at all).

This would be acceptable for me.

Could you please suggest a criteria for "huge" in "max|a[i,*]| is not
very huge".

I have to automate this, because I generate thousand of LPs in an
iteration loop.

Many thanks!

Ali