Hi Antti,
On Mon, 24 Jan 2022, Antti Lehtila wrote:
> Please find below my attempt to explain what's happening there. I
> think the integrality tolerance is parm->tol_int = 1e-5;
> (hard-coded), at least it was so a few years ago.
first, many thanks for checking into this. There is no --tolint
Hi Hartmut,
Please find below my attempt to explain
what's happening there.
I think the integrality tolerance is
parm->tol_int = 1e-5; (hard-coded), at least it was so a few
years ago.
And indeed, the solution for the LP
relaxation
IEEE float has different representations for -0 and +0. Any chance that might
be involved in the computations in some way, even though MIP? What happens
with e != 0?
> On 01/24/2022 3:46 AM Hartmut Henkel wrote:
>
>
> Hi,
>
> here seems to be some one-off error in gmpl/glpk (5.0, debian L
Hello Harmut !
Testing it on https://meimporta.eu/myglpk-ext/ that's GLPK 4.65 plus
several customizations gives this result:
===
v: 131072
v: 0001
v: 131071
e: 0
===
I'm don't see a reason to upgrade right now.
Cheers !
On 24/1/22 10:46, Hartmut Henkel wrote:
Hi,
here s
Hi,
here seems to be some one-off error in gmpl/glpk (5.0, debian Linux):
param v := 131072;
param n := 20; # number of bits
set J := 1..n; # set of all bits
var x{j in J}, binary;
var e, >= 0;
minimize slack: e;
s.t. cs1: sum{j in J} (2**(j - 1) * x[j]) - v <= e;
s.t. cs2: sum{j in J} (2**(j