How to set optimality gap in non-convex quadratic program?Ongoing
I have a non-convex quadratic continuous program and I am willing to stop the software when the optimality gap is 1%. How should I define it? Is there any difference if all variables are continuous vs variables that are mixed integer?
The MIPGap parameter controls the relative optimality gap at which Gurobi stops solving. The default value is 0.0001 (a 0.01% gap). The MIPGap parameter is used for this purpose regardless of whether your non-convex quadratic problem contains integer variables.
For examples of how to set parameters through various APIs, please see the Parameter Examples section of the documentation.0
Is the MIPGap PARAMETER also suitable for the common convex programming or LP?0
Please sign in to leave a comment.