Are there some big changes from version 9.1.1 to 10.0.3?
OngoingHello,
I wanted to ask if there are some major changes between version 9.1.1 and 10.0.3 regarding the modeling of an optimization problem.
When I want to solve a MISOCP in 9.1.1 everything works fine, however, I get a warning that model.cones is deprecated and will be removed in a future version.
When I want to solve the exact same MISOCP in 10.0.3, then my optimization problem has an unbounded objective according to the output log.
I use Matlab and YALMIP. I suppose that some constraint formulation of my model is not valid anymore in the newer versions and therefore leads to an unbounded objective as the constraints won't be recognized as constraints. Which would explain why I get the differences between 9.1.1 and 10.0.3.
Do you have any idea what the issue could be? Or could it be that there is an issue with the intersection between YALMIP and Gurobi?
Thanks a lot!
-
Hi Joschua,
Could you try the latest version (v11.0.0 now)?
Cheers,
David0 -
Hey David,
I tried it with the latest version as well and nothing changed compared to 10.0.3. So still the unbounded objective.
Best,
Joschua
0 -
Hi Joschua,
In that case, this might be an old bug in 9.1.1 that has been fixed with the later versions.
Can you check using another solver?
If not, I can take a look if you can share the MPS file (via a link).Cheers,
David0 -
Hey David,
I tried it with the newest version of Mosek and it returned me approximately the same optimal solution as for Gurobi version 9.1.1.
Best,
Joschua
0
Please sign in to leave a comment.
Comments
4 comments