Get best-known feasible answer after time limit

Question

I am solving a large MIP in Gurobi 6.0. My advisor wants to set a time limit of 12 hours on the problem. I have found that I can set the TimeLimit parameter and that will kill the solver after the alloted time, but I don't know how to retrieve the best feasible solution at that time, just the objective value and optimality gap. Is there a way to get the best feasible solution?

Answer 1

To obtain the best feasible answer so far, you should first verify that there is a feasible solution by checking the Status attribute on your model object, then querying the X attribute on your variable objects. If you have named your variables (with the name parameter on e neat way to do it with the python api is to create a dictionary with the nonzero values.

import math
epsilon = 1e-6
m = grb.Model()
# .....
if m.SolCount > 0:
    solution = {v.varName: v.X for v in m.getVars() if math.abs(v.X) > epsilon}

Answer 2

Have you tried using the Xn attribute, possibly in the combination with the SolutionNumber parameter? Although for your case you should keep the default 0 value of SolutionNumber. You can start from here: http://www.gurobi.com/documentation/6.0/reference-manual/xn.