Re: [Help-glpk] Speeding up GLPK

[Michael Hennebry]

On Fri, 24 Sep 2004, Andrew Makhorin wrote:

> On return lpx_simplex saves the basis information in LPX object, so
> when it is called once again, it starts from the current basis; thus to
> speed up solving you just need to do nothing (however, if you expect the
> current basic solution to be dual feasible, you can enable the dual
> simplex; see the control parameter LPX_K_DUAL). In initial basis all

I seem to remember reading that
degeneracy and approximate arithmetic could make this tricky.
Recomputation after tightening constraints might result in a basis
that is neither dual nor primal feasible.
Similarly for relaxing constraints.

Is this still current information?
Was it ever?

As I handle a lot of M01LPs, this is likely to be significant to me.

> auxiliary variables are basic and all structural ones are non-basic (so
> the basis matrix is unity). The current basis can be changed only by the
> following api routines: lpx_std_basis, lpx_adv_basis, lpx_set_row_stat,
> lpx_set_col_stat, lpx_simplex.

-- 
Mike   address@hidden
"Nothing says it like words if you know how to use them."
                    --  the Professional Organization of English Majors