[Help-glpk] On strict inequalities conversion

[Paulo J. Matos]

On a reply by Andrew Makhorin dated Fri, 9 Aug 2002 20:10:32 +040:
http://lists.gnu.org/archive/html/help-glpk/2002-08/msg00004.html

Andrew says the one can transform a lower strict bound to a lower weak bound as:
eps = 10.0 * lpx_get_real_parm(lp, LPX_K_TOLBND);
new_lb = old_lb + eps * (1.0 + fabs(old_lb));

Why not ?
new_lb = old_lb + lpx_get_real_parm(lp, LPX_K_TOLBND);
?

Still, if I have an upper bound not in a variable but a constraint upper bound:
maximize 0: (which means I don't need to maximize, I need _any_ solution)
sum a_i x_i < b

can I do:
eps = 10.0 * lpx_get_real_parm(lp, LPX_K_TOLBND);
newb = b - eps*(1.0 + fabs(b));

and solve
sum a_i x_i <= newb?

Is this the correct way (best way) to work around this issue?

Cheers,
--
Paulo Jorge Matos - pocm at sat inesc-id pt
Web: http://sat.inesc-id.pt/~pocm
Computer and Software Engineering
INESC-ID - SAT Group