How Gurobi's concurrent MIP solver works
AnsweredHi!
I am using your concurrent MIP solver in PyCharm to solve my fleet size and mix problem for the installation process of floating offshore wind. When I increase the number of turbines to be installed and vessels the model can choose from the problem increase with many variables due to a lot of for loops containing these and the operations they are going to perform.
What I wonder is if I have understood how the concurrent MIP solver works for a minimization problem. Does it use primal and dual simplex to figure out the lower bound, which it uses for the further search in the branch and cut algorithm?
How do the branch and cut algorithm work? Because I have seen that the solver works faster if I increase the number of vessels from 10 to 22 for 11 turbines. In my problem, I need at least two vessels per turbine to execute the operations. Can this be the reason why it works faster for a larger problem with 22 vessels? Is it more difficult to figure out how to utilize a fleet of 10 vessels rather than 22 vessels, because for the 10 vessels some of them have to be used for more than one turbine?
Best regards,
Sebastian

Hi Sebastian,
Please see these sections in our documentation for further information about how Gurobi solves MIPs concurrently:
It is often very hard to explain why a certain model is more difficult to solve than another. It may happen that the additional degrees of freedom in the larger model makes it easier to find a feasible solution and better guides the solver to optimality.
Cheers,
Matthias
Please sign in to leave a comment.
Comments
1 comment