Re: [Help-glpk] small coefficient in glpk?

[Andrew Makhorin]

> I'm using glpk 4.8 (via Java JNI
> http://bjoern.dapnet.de/glpk/index.htm) as a feasibility solver. One
> of my constraint is that the sum of structural variables = 1.

> When some coefficients are very small, e.g. 1E-6, then the constraint
> that the sum of structural variables = 1 is not enforced. I got all of
> them with 0. If I ignore those small coefficients (<= 1E-6), the
> constraint that the sum of structural variables = 1 is satisfied. But
> I'm not sure if this will affect the accuracy of my program.

> Does anyone know if glpk has some precision lower bound or this is a bug?

Your question is unclear. Could you give an example of the constraint
and explain what is wrong?

(Note that the most recent version of glpk is 4.32.)