[Bug-glpk] Re: pb with -DBL_MAX

[Andrew Makhorin]

> I think I found a bug in simplex algorithm.
> When I fix some costs to -DBL_MAX in the objective function, glpk never 
> goes out from simplex...
> it just runs like in an eternal cycle.

I'm not sure why you fix some costs to -DBL_MAX. Using huge objective
coefficients is a bad idea because this involves serious numeric troubles.
Probably you should reformulate your problem.