PIPS-NLP
|
Classes | |
class | NlpGenLinsys |
class | NlpGenResiduals |
class | NlpGenSparseLinsys |
class | NlpGenVars |
class | sResiduals |
OOQP's default general problem formulation:
minimize f(x) ; subject to A(x) = b ; <-- dual y clow <= C(x) <= cupp ; xlow <= x <= xupp ;
The general linear equality constraints must have either an upper or lower bound, but need not have both bounds. The variables may have no bounds; an upper bound; a lower bound or both an upper and lower bound.
add slacks as: t: slack for lower bound: C(x)-t=clow u: slack for lower bound: C(x)+u=cupp lambda: the dual var for the lower bound constraint C(x)-t-clow=0 pi: dual var for the upper bound C(x)+u-cupp=0
Or: add slacks as: s: C(x)=s; t: slack for lower bound: s-t=clow u: slack for lower bound: s+u=cupp lambda: the dual var for the lower bound constraint s-t-clow=0 pi: dual var for the upper bound s+u-cupp=0
xlow <= x <= xupp => x-v=xlow, x+w=xupp v: slack for lower bound: x-v=xlow w: slack for lower bound: x+w=xupp gamma: dual var for the lower bound phi: dual var for the upper bound