[Help-glpk] glp_ios_heur_sol (sorry for the first mail in the wrong format)

[Anne-Laurence Putz]

Dear support,

I would like to fix all variables of a MILP problem before launching
optimization (to help the branch&cut solver with an upper bound, for
instance).

Those values are taken either from a custom heuristic of my own or from a
previously launched MIP optimization.

Anyway, I don't understand how to do such a thing with GLPK 4.38.

I would like to use glp_ios_heur_sol(), but it doesn't seem to be the method
I'm searching for (I don't understand what is the "primal heuristic"  
that we have to launch before using glp_ios_heur_sol()).

But I don't know how to do it differently.

You gave an answer on thread "Setting initial integer solution in GLPSOL" (see
http://www.nabble.com/Setting-initial-integer-solution-in-GLPSOL-to21005682.
html#a21008418).

If I've correctly understood, you said that one have to patch GLPK to do so
? Because there isn't any parameter in glp_iocp structure indicating that
the current solution stored comes from a heuristically found solution.

In fact, I would like to call a glp_initial_sol(lpProb*, const double
values[]) which would fix all the MILP variables with the 'values' array and
check for feasibility (and ignore the initial solution if it's infeasible).
Is there any way to do it with GLPK 4.38 ?

Best regards,

Anne-Laurence Putz

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.