A linear program is an optimization problem with
linear cost function,
equality constraints and non-negative unknowns
Usually, it is assumed that the
matrix
has full rank . This can always
be achieved by removing redundant constraints.
General linear constraints
for unknowns can be brought into the
above standard form by introducing auxiliary variables.
We write as the difference of two non-negative vectors,
and introduce non-negative slack variables to remove
the inequality constraints. More precisely,
is replaced by
where
.
(Authors: Höllig/Pfeil/Walter)
|
automatically generated
4/24/2007 |