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
The conic solvers don't support any variable constraints. So when implementing copy_to from their OptimizerCache, they will all need to check that no constraint variable is set. LP solvers supports some of them but not all so they will need to check that the ones they support are not set.
One way to improve the situation would be to allow GenericModel to have different type of variable containers.
The text was updated successfully, but these errors were encountered:
It would but I feel it's best too just make GenericModel modular.
Now it's just a combination of three submodels: one for the objective, one for the scalar variables constraints and one for the rest. It's just dispatching calls to the correct submodel, it's not specific at all to what is used as submodel. Allowing it to have any submodel would enable any combination of models. That also allows supporting other kind of SingleVariable constraints and other function as objective.
The conic solvers don't support any variable constraints. So when implementing
copy_to
from theirOptimizerCache
, they will all need to check that no constraint variable is set. LP solvers supports some of them but not all so they will need to check that the ones they support are not set.One way to improve the situation would be to allow
GenericModel
to have different type of variable containers.The text was updated successfully, but these errors were encountered: