How to get the relative MIP optimality gap after timelimit is met?

Question

I'm working on a large scale MIP. So I have to set the time limit to a reasonable value. But the problem is that after the timelimit is met, I don't know how to evaluate the solution, in another words, the gap between lower and upper bounds. Yes, the gap will be shown on screen. But if there is a way to get its value with Python API? So I can output it or do some following statistcs. I have to solve more than 10 MIP at once, it's hard to find the gaps one by one on screen. Thx for your help!

Answer 1

Your tags indicate that you are using Gurobi with the Python API to solve your models.

To retrieve the relative MIP gap, you can query the MIPGap attribute (see http://www.gurobi.com/documentation/6.5/refman/mipgap.html)

Python example:

from gurobipy import *

model = read("model.mps")
model.params.TimeLimit = 100
model.optimize()
print("Final MIP gap value: %f" % model.MIPGap)