Is the original model or presolved model dictating the algorithm?

Answered

Elina Joksch

May 02, 2022 16:10

• 
• 
• 

Dear Community, 

my original model is continous, but after the presolve I have an MILP. Is in this case the algorithm for continous models used (Method = 3, concurrent) or Branch and bound for MILP?

Model has 2 quadratic objective terms
Model has 2 quadratic constraints
Model has 3 general constraints
Variable types: 14 continuous, 0 integer (0 binary)
Coefficient statistics:
  Matrix range     [2e-03, 1e+00]
  QMatrix range    [1e+00, 1e+00]
  QLMatrix range   [1e+00, 1e+00]
  Objective range  [6e+02, 6e+02]
  QObjective range [1e+03, 1e+03]
  Bounds range     [1e+00, 5e+02]
  RHS range        [1e+00, 5e+02]
  QRHS range       [1e+00, 1e+00]
Presolve added 60 rows and 1324 columns
Presolve time: 0.03s
Presolved: 72 rows, 1339 columns, 24882 nonzeros
Presolved model has 2 bilinear constraint(s)
Variable types: 1311 continuous, 28 integer (5 binary)

 

Thank you very much in advance!

Greetings,

Elina

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  November 16, 2025 23:38

  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.
• 

  Jonasz Staszek

  • Community Moderator

  May 02, 2022 18:37

  • 
  • 
  • 

  Hi Elina,

  AFAIK it is the presolved model. Perhaps someone from the Gurobi team can confirm this?

  The log suggests that presolve added both constraints and variables. This could be due to the linearization of the quadratic and general constraints you included. This usually simplifies the model and allows for its faster solution.

  To compare the two models - original and presolved - you could try:

  model1 = model.presolve()
  model1.write("presolved_model.lp")

  Hope this helps.

  Best regards,
  Jonasz

  1
• 

  Simon Bowly

  • Gurobi Staff

  May 03, 2022 02:30

  Yes, the presolved model structure dictates the algorithm that is used. If the presolved model is an MILP, then in general branch and bound is needed to guarantee a result.

  Most likely, one of the general constraints required an auxiliary binary variable to be added to the model in order to handle it. So, the model is not really continuous.

  0
• 

  Elina Joksch

  May 19, 2022 12:27

  • 
  • 
  • 

  Thank you very much! This helped me a lot!

  0

Post is closed for comments.