I have just upgraded to Gurobi v. 10, which lead to that my previously built algorithm (based on a MILP) no longer works in some cases. I debugged it thoroughly, and it seems that v. 10 of Gurobi behaves differently (and worse) than its predecessor for a small test case. What happens is that the preprocessing step removes all lines, and the variables that comes out to my understanding do not have correct values. I have exported the model, but I don't see a way to attach files here. I realize that it could be the case that there is just one solution, but shouldn't I then get the same value of the variables as with the previous version? Any suggestions?
Please sign in to leave a comment.