MILP model feasibility check speed
AnsweredDear Gurobi Community,
How should I set the Gurobi parameters for just checking the feasibility of milp models, as fast as possible?
I have tried my models on NEOS and one another server (on my university site) with proper hardware specification (plenty of RAM and number of threads). Solving the models on NEOS take around a few minutes, while it takes more than a day for some models on the university server. I guess it might be due to different Gurobi parameters. However, I do not know the best parameters value for just checking the feasibility. Any idea will be so appreciated.
Thanks in advance.
-
Official comment
This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try our AI Gurobot?. -
Hi,
Since you are interested only in feasibility of a given model, I assume that your objective function equals 0. This means that Gurobi terminates as soon as a single feasible solution is found. You should always use the latest Gurobi version. As for parameters, you should have a look at the documentation of Most important parameters for MIPs. In particular the no relaxation heuristic may help. Other than that, you should experiment with the parameters, MIPFocus, Heuristics, ImproveStartNodes, ImproveStartTime, Presolve, ZeroObjNodes, MinRelNodes. On a side note, you can limit the number of solutions to be found by SolutionLimit.
Best regards,
Jaromił0
Post is closed for comments.
Comments
2 comments