Providing a feasible solution right at the start can be very helpful to improve the time to optimality. You can do this either by using the Start variable attribute or by reading an MST or SOL file on the command line via the InputFile parameter.

We recommend only setting the binary and integer variable values and let Gurobi figure out the continuous solution values on its own. This can help generate a feasible solution when a full start vector might lead to some numerical violations and is not accepted.

Please make sure to check the MIP Start documentation for further information.

There is a related option that you might want to try: Variable hints. These values serve as guidance throughout the solving process. So when you know that a variable is very likely to have a certain value in the solution you can set this variable hint and help Gurobi find a solution.

## Comments

0 comments

Article is closed for comments.