Re: [Help-glpk] non-official updated version of glpk (4.63 pre-release)

[Chris Matrakidis]

Andrew,

I run into this behaviour with the released version as well (with a different MIP):

Warning: dual simplex failed due to excessive numerical instability

and later on it seems to get stuck in a sequence like:
*279500: obj =   2.196544071e+08 inf =   1.239e-13 (2)
*280000: obj =   2.196544071e+08 inf =   1.239e-13 (2)
*280500: obj =   2.196544071e+08 inf =   1.239e-13 (2)
...
*9366000: obj =   2.196544071e+08 inf =   1.239e-13 (2)
*9366500: obj =   2.196544071e+08 inf =   1.239e-13 (2)
*9367000: obj =   2.196544071e+08 inf =   1.239e-13 (2)
and so on.

In this case I see the following:

Warning: numerical instability (dual simplex, phase II)
...
Warning: numerical instability (dual simplex, phase II)
Warning: dual simplex failed due to excessive numerical instability
*4718105: obj =  -2.406747734e+06 inf =   1.988e-08 (1)
*4999119: obj =  -2.406747734e+06 inf =   1.988e-08 (1)
...
*156100559: obj =  -2.406747734e+06 inf =   1.988e-08 (1)
*156388395: obj =  -2.406747734e+06 inf =   1.988e-08 (1)

and so on.

I did some additional checking and it seems that spx_chuzc_sel finds an eligible variable with d[j] ~3 times the eps value but no progress is made, ending up cycling between two variables.

Is this enough information to go on, or do you want me to check any other variables or provide additional details?

Best Regards,

Chris Matrakidis