[Coin-discuss] Access to dual variable corresponding to a newrow

[Matthew Saltzman]

On Wed, 5 Dec 2001, I wrote:

> According to CPLEX docs, CPXdelrows() and CPXdelsetrows() do compress the
> deleted rows out of the index list.  The implication is clearly that added
> rows are added at the bottom (although the CPXaddrows() docs to not state
> this explicitly.

FWIW, I think if we ever encountered a solver that didn't behave this way,
we'd be obligated to do any different index tracking within the OSI.
Whatever the solver does, the OSI needs to present a consistent API--or
possibly alternative ones, but not solver-dependent--any scheme we support
must work with any solver we support.

-- 
		Matthew Saltzman

Clemson University Math Sciences          Editor, INFORMS Online
mjs at clemson.edu                           iol_editor at mail.informs.org
http://www.math.clemson.edu/~mjs          http://www.informs.org