Very slow hierarchical objective optimization

回答済み

Alexia Payan

2021年03月25日 17:48

• 
• 
• 

I a considering a hierarchical objective optimization. 

The first objective already takes a long time to "solve" to a given gap, say 3%.

When adding the second objective, the runtime is increased by a lot and sometimes, the gap get stuck for a very very long time at a given value and does not improve much over very long periods of time.

Is that expected?

 

Also, what is the meaning of the gap for the second objective function assuming we are willing to degrade by 10% (RelTol = 0.1) the first objective?

Can we expect to go to low gaps, say 3% on the second objective by degrading the first one by 10%?

 

Thank you.

0

• 正式なコメント

  

  Simranjit Kaur

  • Gurobi Staff

  2025年06月23日 02:57 編集済み

  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

  2021年03月26日 08:55

  Hi Alexia,

  If already solving for 1 objective is hard, then solving for the second one is expected to be harder, because the first objective is added as an additional constraint to the problem.

  Can we expect to go to low gaps, say 3% on the second objective by degrading the first one by 10%?

  Yes, this is what we can expect. You can find all details in the Working With Multiple Objectives documentation.

  In general, it might be helpful if you first try to find good parameters for the first objective only. If the performance for the first objective is improved, it may positively affect the second objective as well. You can have a look at Most important parameters section and/or try the Gurobi Parameter Tuning Tool.

  Best regards,
  Jaromił

  0

投稿コメントは受け付けていません。