Skip to main content

Barrier Method Sub-Optimal Termination



1 comment

  • Jaromił Najman
    Gurobi Staff Gurobi Staff

    Sub-optimal termination is most often caused by numerical issues during the solution process, please refer to our Guidelines for Numerical Issues for more information. You should try running your model with the BarHomogeneous parameter set to 1 to possibly improve numerical behavior of the algorithm.

    Gurobi chooses the last best iterate which has a best combination of low primal and dual violations while still having a rather low complementarity value. This choice is made to avoid returning an infeasible solution as would be the case if Gurobi just returned the last iterate which has a primal violation of  2.52e-02 and dual violation of 2.05e-03.

    Best regards, 


Please sign in to leave a comment.