Tobias Achterberg

Hi Austin, yes, your observation is correct. The fix in Gurobi 9.5 just means that those solutions are discarded from the solution pool. It does not mean that they are already pruned in the search ...

Thanks. I found the issue. Even if you disable cuts, we are allowing cuts to be generated in a cut loop that runs in parallel to the root node solve. As a workaround, you could instead set the CutP...

Hi Ellen, this is strange. I would like to investigate. Do you have an mps file with which you can reproduce the issue using the Gurobi command line tool? On Linux, you would call gurobi_cl Cuts=0 ...

Having an empty callback should not deteriorate performance much. If you are using Python, then of course a callback means that we need to always step into the Python interpreter, but even this sho...

It is an interesting coincidence that you are running into this issue now, because I am currently trying to improve this behavior of Gurobi. I think that I found a nice way to do so and will pick u...

Hi Thomas, Jaromils suggestions are of course perfectly fine. But I wonder why it is bad for you to just set a time limit that fits to the difficult scenarios for the overall multi-scenario solve. ...

Thanks for pointing out this bug in our documentation. This is still from Gurobi versions prior to 9.0, where an equality Q constraint would automatically lead to a "non-PSD" error (quadratic equat...

Yes. You really should take a look at the optimization status code in d.status. If it is GRB.INFEASIBLE, then the model is infeasible. If it is GRB.UNBOUNDED, then the model has an unbounded ray (b...

The KeyError is not from Gurobi. I guess this comes directly from Python and is triggered in the gurobipy module. Could it be that there is some j in M and i in N for which S[i,j] or v[j] does not ...

What is d.status? If it is GRB.INF_OR_UNBD or GRB.UNBOUNDED, then you won't be able to compute an IIS. You will only get an IIS for a provably infeasible model. If the status is GRB.INF_OR_UNBD, th...