Re: [Help-glpk] Presolver not found for interior point algorithms

[Andrew Makhorin]

> Generally, the preprocessor steps for the simplex algorithm work
> very well for interior point algorithms; I see no reason why you
> should not naturally allow all of those tricks to be used for either
> solver.

> The only solver-specific preprocessor operation that I know is the
> splitting up of dense columns needed by the interior point algoritms.
> But I do not see any reason not to give both solvers all the "simplex"
> preprocessing tricks.

Generally, yes. However, there are some technical differences on
recovering (i.e. postprocessing) basic and interior-point solutions
obtained for the preprocessed problem.