Is it possible to retrieve the best (lower) bound of a MILP problem that did not reach the optimal solution?

Answered

Efrain Ruiz

• March 22, 2021 16:55

I have a MILP model, which I solve using time_limit. Many times the optimal solution is not reached. I can retrieve the incumbent solution, but I want to retrieve the fractional solution of the best (lower) bound generated during the Branch and Bound process. Is there a way to get it? I am using C++ API.

0

• 

  Jaromił Najman

  Gurobi Staff

  • March 22, 2021 19:57

  Hi Efrain,

  You can access the best bound via the ObjBound attribute.

  Best regards,
  Jaromił

  0
• 

  Efrain Ruiz

  • March 22, 2021 21:18

  

  

  

  Sorry, I was not clear enough I want to retrieve the value of the variables of the best lower bound solution, not the value of the objective function.

  0
• 

  Efrain Ruiz

  • March 23, 2021 03:55

  

  

  

  I have just found how to do it using a callback. You have to use

  if (where ==GRB_CB_MIPNODE && getIntInfo(GRB_CB_MIPNODE_STATUS) == GRB_OPTIMAL)

  to find out if the node is optimal for the LR of the problem

  then use

  getNodeRel () to obtain the value on the variables. Thank you Jaromil, for your help

   

  0

Please sign in to leave a comment.