[Help-glpk] Re: Fwd: Degeneracy problem

[Andrew Makhorin]

On Tue, 2011-01-11 at 22:04 -0800, Abbas Ebadi wrote:
> ---------- Forwarded message ----------
> From: Abbas Ebadi <address@hidden>
> Date: Mon, 10 Jan 2011 04:12:55 -0800
> Subject: Degeneracy problem
> To: address@hidden
> 
> Dear Mr or Madam,
> 
> I am using the GLPK solver for solving an LP model in GAMS. At primary
> iterations, it improves the value of objective function rapidly but,
> then, solution process continues so long and very slow progress
> happens. I was wondering how I can prevent the solver to continue
> solving process when such cases happen.
> 
> Kind Regards,
> Abbas Ebadi
> 

This may happen if your lp instance is highly degenerate. You can try to
increase the feasibility tolerance (tol_bnd) or use the dual simplex (by
default the primal simplex is used). Does the glpk simplex solver solve
the instance to optimality, or it stalls with no progress?