[Scip] SCIPchgVarObj and SCIPaddObjoffset

[Tobias Achterberg]

Gerald Gamrath wrote:
> The method SCIPaddObjoffset is meant to be used only in the presolving
> stage, so you are not allowed to call it at the points you tried. I
> don't know why this is the case, it would probabaly be possible to make
> it available also in the problem creation stage.

I am pretty sure that one can make the method available also in problem creation stage. 
Not having done so is just the result of our strategy to only make methods available in 
those stages where we think they are useful (and are therefore tested).

> However, I think there are two easy solutions for you: Either you create
> an "obj" variable that is fixed to zero and gets the offset as objective
> coefficient, or you just store the dual value of the convexity
> constraint in your pricer and subtract it from your subscip's optimal
> value afterwards.

Of course, you need to fix the dummy "obj" variable to 1.


Tobias