Adding lazy cuts using callback in LP

Answered

Yingqiu Zhang

• April 29, 2021 05:28

Is it possible to add lazy cuts using callback feature in a linear program? I am currently solving a stochastic linear program using Benders decomposition. But I observed that Gurobi can only retrieve solutions for MIP. Is it possible to retrieve the current solution of an LP when an optimal solution is found?

Thanks!

0

• Official comment

  

  Simranjit Kaur

  Gurobi Staff

  • December 06, 2024 00:32

  This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try our AI Gurobot?.
• 

  Jaromił Najman

  Gurobi Staff

  • April 29, 2021 07:33

  Hi,

  Adding lazy cuts during the solution of a linear program does not make sense, because there are no intermediate solutions to be cut off. There is no B&B tree present and no subproblems have to be solved. For linear programs the procedure you describe should be:

  1. Solve LP to optimality
  2. Get optimal solution and analyze it
  3. Add new cut to LP and re-optimize

  This has the advantage that Gurobi can re-use information from a previous solve and warm-start subsequent optimization processes improving the performance.

  Best regards,
  Jaromił

  0
• 

  Yingqiu Zhang

  • April 29, 2021 18:56

  

  

  

  Hi,

  Thanks for the answer! As for the aforementioned procedure, can I still use the callback feature? Or I can only reoptimize it and add cut iteratively?

   

  Best,

  Yingqiu

  0
• 

  Jaromił Najman

  Gurobi Staff

  • May 19, 2021 13:07
  • Edited

  Hi Yingqiu

  Please excuse the late response.

  Lazy and user cuts are not available for LPs. Thus you have to re-optimize and add cuts iteratively when working with LPs.

  Best regards,
  Jaromił

  0

Post is closed for comments.