You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using OsiXpr, I need to be able to use the Newton barrier solver. It's just that much more efficient with my application.
The present behaviour is to use automatic solver selection for initialSolve() and to use dual simplex for resolve().
There is no means of changing this behaviour in OsiSolverInterface or OsiXprSolverInterface. There is no access to the XPRSprob member _prob. Even if there were, this would not help when calling resolve(). Likewise, there are no parameter(s) in OsiSolverInterface to give this kind of control over the solver behaviour.
At present, I'm forced to produce my own version of OsiXprSolverInterface. It would be neater to avoid this, say, by having a parameter in OsiSolverInterface that allows the selection of a preferred method whenever there is a choice available (say interior point/primal simplex/dual simplex).
Thanks, Dan
The text was updated successfully, but these errors were encountered:
Issue created by migration from Trac.
Original creator: dan.gordon
Original creation time: 2006-08-25 04:18:10
Assignee: @mjsaltzman
Version:
When using OsiXpr, I need to be able to use the Newton barrier solver. It's just that much more efficient with my application.
The present behaviour is to use automatic solver selection for
initialSolve()
and to use dual simplex forresolve()
.There is no means of changing this behaviour in OsiSolverInterface or OsiXprSolverInterface. There is no access to the
XPRSprob
member_prob
. Even if there were, this would not help when callingresolve()
. Likewise, there are no parameter(s) in OsiSolverInterface to give this kind of control over the solver behaviour.At present, I'm forced to produce my own version of OsiXprSolverInterface. It would be neater to avoid this, say, by having a parameter in OsiSolverInterface that allows the selection of a preferred method whenever there is a choice available (say interior point/primal simplex/dual simplex).
Thanks, Dan
The text was updated successfully, but these errors were encountered: