Getting the dual variable values when the model status is 4(INF_OR_UNBD)

Answered

Iason Liagkas

December 21, 2023 22:35

• 
• 
• 

Hello guys, I am running the benders algo, When my subproblem is feasible I am using the following line to get the dual variable value: 

gp.quicksum(BigM*deltars2kE[r,s,k]*subproblem.getConstrByName("UE_EV_2[E,{},{},{}]".format(r,s,k)).Pi for r,s in drs.keys() for k in K)

And it works. But when my subproblem becomes infeasible or unbounded and I want to take the dual variable values I am using the exact same code but with .FarkasDual instead of .Pi 

gp.quicksum(BigM*deltars2kE[r,s,k]*subproblem.getConstrByName("UE_EV_2[E,{},{},{}]".format(r,s,k)).FarkasDual for r,s in drs.keys() for k in K)

Then I get the following error:
AttributeError: Unable to retrieve attribute 'FarkasDual'.
Am I doing something wrong? Are there any specific conditions that the attribute FarkasDual works? 

 

Related to

• Sensitivity

0

• 

  Eli Towle

  • Gurobi Staff

  December 21, 2023 23:41

  From the FarkasDual documentation:

  These attributes are only available when parameter InfUnbdInfo is set to 1.

  Have you set the InfUnbdInfo parameter to 1?

  0
• 

  Iason Liagkas

  December 27, 2023 07:01

  • 
  • 
  • 

  Thank you Eli, that was the problem, enjoy your holidays! 
  
  Kind regards
  Iason

  0

Please sign in to leave a comment.