Strategies for solving a problem with low number of decision variables?

回答済み

s

2022年08月05日 18:16

• 
• 
• 

Newbie here.

 

So I have an optimization model that has a very low number of decision variables compared to the total number of variables in the model, which consist of both continuous and binary. All decision variables are binary so we could even use brute force to solve the same problem if the number of possible combinations is manageable!

However, I am wondering if a tuning strategy exists when all decision variables are binary and the number of binary variables is low compared to the total number of variables in the model?

 

Thanks.

0

• 正式なコメント

  

  Simranjit Kaur

  • Gurobi Staff

  2025年11月16日 23:31

  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 try Gurobot, our chatbot interface offering instant, expert-level support.
• 

  Matthias Miltenberger

  • Gurobi Staff

  2022年08月08日 10:03

  Short answer: No, you should never use brute force. The number of possibilities is growing exponentially with the number of decision variables. 20 binary variables already result in more than a million combinations.

  Gurobi will handle those combinatorial problems way more efficiently than brute force ever could. In almost every practical problem there are structures and inter-dependencies between the variables that can be exploited to reduce the complexity of the problem and make it solve faster.

  I hope that answers your question.

  Cheers,
  Matthias

  -1
• 

  s

  2022年08月08日 14:16

  • 
  • 
  • 

  Well I am well aware that Gurobi can do that, but my question is: what is the tuning strategy for such a problem (e.g. what parameter to tune to improve the performance OR what is the modelling strategy) to improve the speed of solution?

  0
• 

  Matthias Miltenberger

  • Gurobi Staff

  2022年08月08日 14:56

  The amount of integer or binary variables is not enough to say something about which setting might improve the solver. Vastly different models can have the same low number of these variables but still react differently to various tuning strategies.

  Cheers,
  Matthias

  0

投稿コメントは受け付けていません。