Retrieve relative gap and dual objective value

ユーザーの入力を待っています。

jake roth

2023年03月07日 16:37

• 
• 
• 

I am running a barrier method to solve an LP and a QP and would like to retrieve

1. final residuals: primal infeasibility, dual infeasibility, relative objective value gap
2. final dual objective value

For (1), is this: MaxVio, DualVio, ComplVio?

For (2), is there an attribute for this?

 

関連

• Tolerances
• Attribute

1

• 正式なコメント

  

  Simranjit Kaur

  • Gurobi Staff

  2026年11月13日 01:47

  This post is more than two 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.
• 

  Mario Ruthmair

  • Gurobi Staff

  2023年03月13日 09:19

  Hi Jake,

  (1) There are only solution quality attributes related to the model, not to particular variables or constraints. They give you the maximal violations over the whole solution:

  • BoundVio, ConstrVio, ConstrResidual give you the primal violations
  • DualVio, DualResidual give you the dual violations
  • ComplVio gives you the complementarity of the current primal/dual solution pair (which should be zero)
  • MaxVio is the maximum over all the above.

  (2) You can query the dual objective value via model attribute ObjBound.

  Best regards,
  Mario

  1
• 

  jake roth

  2023年07月25日 19:52 編集済み

  • 
  • 
  • 

  Thanks Mario! When I query this field for an LP solved with barrier, I find that I obtain an error (1005: DATA_NOT_AVAILABLE) though the optimal objective is nonzero and the solver terminates successfully with the correct solution. Is there a field/attribute that contains the final dual residual value from the barrier log?

  0
• 

  Mario Ruthmair

  • Gurobi Staff

  2023年07月31日 10:56

  I just solved an LP with barrier and was able to query successfully the attribute "DualResidual" after the optimization finished. Could it be that your final state was not "Optimal"?

   

  1

投稿コメントは受け付けていません。