MIP-Heuristic as a MIP-start

Answered

A.Omidi

• March 30, 2020 10:28

Dear support team,

 

I'm trying to solve a scheduling problem using Gurobi 9. The model is solved and the optimal solution is found. Now, I would like to use a heuristic (which an instance could be found here), as a MIP-start to speed up solving the problem. I'm using Java to write the model. Unfortunately, I cannot find any related document about the greedy heuristic on the Gurobi host. 

I was wondering if, is there any example or document in which, how can I implement this method using Java?

 

Regards

A. Omidi

0

• 

  Eli Towle

  Gurobi Staff

  • March 30, 2020 23:39

  Hi,

  You can specify a MIP start vector by setting the Start variable attributes. An example of how this is done in Java can be found in the Facility.java example, part of the examples library.

  To be clear, the greedy heuristic function (greedy_heur) referenced in that PDF presentation is not an actual Gurobi function. It is meant to be a placeholder for a user-defined function which populates the greedy_x dictionary, which is then used to assign the Start values for the model variables. Thus, if you want to provide a MIP start to Gurobi using a greedy heuristic, you will have to implement the actual heuristic yourself.

  Thanks!

  Eli

  0

Please sign in to leave a comment.