Solution improves with additional constraints

Awaiting user input

Henrik Vassbotten

• May 07, 2024 16:07

Hi! 

We're using Gurobi's solver for a university project, but have run into a problem. 
We are adding constraints sequentially to a max problem (as part of L-shaped method), but the objective value of the problem improves in some iterations, when it should decrease or remain the same each iteration. The increase is sometimes quite substantial (up to 15%). 

For instance, in a max problem:
We solve the problem -> Objective value = 40 000 000
We then add one more constraint to the same problem -> Objective value = 42 000 000

We have printed the model per iteration, and can see that the only changes to the model are the added constraints, otherwise they are identical. 

Can you help us understand why the objective value sometimes increases?

 

Kind regards

Related to

• Bender's

1

• 

  Jaromił Najman

  Gurobi Staff

  • May 08, 2024 07:28
  • Edited

  Hi Henrik,

  Could you please share the two models you mentioned? Note that uploading files in the Community Forum is not possible but we discuss an alternative in Posting to the Community Forum.

  Additionally, could you please share the log output for both optimization runs?

  Best regards, 
  Jaromił

  0

Please sign in to leave a comment.