Re: [Help-glpk] MIP Solvers (i.e CBC, CPLEX, GLPK, GUROBI, LPSOLVE, SCIPC, SCIPL, SCIPS and XPRESS) Benchmark

[Andrew Makhorin]

> > Looks like your instance is maximization while mps always assumes
> > minimization. You need either to change the objective sign in the
> > mathprog model or specify --max option for glpsol.
> 
> I don't understand this statement.  Where do you "specify -max option
> for glpsol"?
> 
> e.g. glpsol --math TimberHarvestModel_0025p.mod --wfreemps
> TimberHarvestM_freemps.mps
> 
> The objective function of the MatProg model is already maximize.
> 
> e.g.
> 
> maximize Discounted_Net_Value:
>     sum {s in STAND, p in PERIOD} Stand_Cut_Candidates[s,p] *
> NPV_Calc[s,p] * FStand[s,p];
> 

When you pass mps to glpsol, the optimization direction is minimization
by default, so you need to specify the --max option to tell glpsol to
maximize rather than minimize.