Obtain the current optimization result of Gurobi under jupyter-notebook

Answered

Di Lv

• June 24, 2024 07:40

I'm using Gurobi to solve a mixed-integer linear programming problem in a web-based jupyter notebook. The optimality gap is currently about 3%, and I'd like to obtain the current solution and stop the further computation. How can I deal with that

Related to

• Termination

0

• 

  Riley Clement

  Gurobi Staff

  • June 24, 2024 11:17

  Hi Di,

  Jupyter Notebooks have different implementations, depending on where they are hosted, but typically there should be an "interrupt" or "stop" button in the interface.

  
  If you use this then Gurobi will stop at the next available point (it may not be immediate).  You can then query the solution and objective value as per usual.

  - Riley

  0
• 

  Di Lv

  • June 25, 2024 09:37

  

  

  Thank you very much！ I have clicked the "interrupt" button, but it doesn't react for a while(about 5 minutes). And the jupyter is still running (with *) on the left. Is this normal?

  0
• 

  Marika Karbstein

  Gurobi Staff

  • June 25, 2024 15:27

  I have the same experience. And I did not find any solution for this.
  For me, the only option is to set termination parameters in the code when running Gurobi in Jupyter notebook, i.e., set TimeLimit and/or MIPGap.

  0
• 

  Riley Clement

  Gurobi Staff

  • June 26, 2024 23:57

  Note that this is not Gurobi specific - failure to reliably interrupt the Jupyter kernel is a long standing issue:
  https://github.com/ipython/ipython/issues/3400

  0

Please sign in to leave a comment.