Since it took quite a long time to solve a model on a 4-core machine (one week), I also started it on a 24-core machine and set the ConcurrentMIP parameter to 4. Apart from the time advantage, Gurobi also found a slightly better solution this way! I wondered why. As far as I understood, setting ConcurrentMIP=4 means that Gurobi explores four completely different parts of the branch-and-bound tree simultaneously, using 6 threads each? Is it possible that the optimizer finds a parameter setting in the process that is a bit better - like the global optimum rather than a local one (even if it is non-determenistic)?
Please sign in to leave a comment.