Key error

Answered

Hernani Delgado Chantre

• April 18, 2021 16:58
• Edited

Hello there, I wonder if you could help me on the following contraint

This is how I understand this constraint can be translated on gurobi but I am getting a match error

Thank you in advance,
Hernani

0

• 

  Jaromił Najman

  Gurobi Staff

  • April 19, 2021 09:32

  Hi Hernani,

  What is the exact error you are getting? Could you post the code you have used so far to implement this constraint?

  Best regards,
  Jaromił

  0
• 

  Hernani Delgado Chantre

  • April 19, 2021 10:31

  

  

  

  Hi Jaromil,

  My apologize I though I copy the code to the post. Here is how I coded the constraint,

  model.addConstrs(
      quicksum(
          x[u, v, i, 0] * math.log(1 - r_u[u]) + x[m, v, i, k] * math.log(1 - r_u[m])
          for m in U
          for k in range(K)
          if u != m
      )
      <= math.log(1 - R_v[v])
      for v in V
      for i in range(C_u[u])
  )

  Regards,
  Hernani

  0
• 

  Matthias Miltenberger

  Gurobi Staff

  • April 19, 2021 16:59

  Hi Hernani,

  I suppose \(r_u\) refers to a Gurobi variable. In this case, you cannot simply apply the logarithm to this variable and then use the result in a constraint. You first need to add auxiliary variables to capture these logs and then use those variables in the final constraint.

  I hope that helps.

  Cheers,
  Matthias

  0
• 

  Hernani Delgado Chantre

  • April 20, 2021 09:50

  

  

  

  Hi Matthias,

  Sorry I did not make it clear but the decision variables are x_uvik, r_u[u] and r_u[m] and R_v[v] are constants.

  Regards,

  Hernani

   

  0
• 

  Hernani Delgado Chantre

  • April 27, 2021 10:41

  

  

  

  Dear all,

  I have this constraint (2) where x_uvik is binary decision variable and r_u, r_m and R_v are parameters to the problem. 

  I coded the problem as follow: But when I run my program with this constraint I get the model is unfeasible unable to retrieve attribute X. Could you please point me the right direction to solve this.

  for v in V:

          constraint = LinExpr()

          constraint1 = LinExpr()

          for u in U:

              for i in range(C_u[u]):

                  constraint.add(x[u,v,i,0]*math.log(1-r_u[u]))

                  for m in U:

                      for j in range(C_u[m]):

                          if (u!=m):

                              for k in range(K):

                                  constraint1.add(x[m,v,j,k]*math.log(1-r_u[m]))

          model.addConstr( constraint + constraint1 <=math.log(1-R_v[v]))

  Regards,

  Hernani

  0
• 

  Matthias Miltenberger

  Gurobi Staff

  • April 27, 2021 12:11

  Hi Hernani,

  You could try to solve the IIS of this model to figure out what might be the cause of the infeasibility:
  Model.computeIIS() (gurobi.com)

  Cheers,
  Matthias

  0
• 

  Hernani Delgado Chantre

  • April 27, 2021 12:16

  

  

  

  Hi Matthias,

  Thank you for quick response, regarding to the code of the constraint for your experience it seems correct or there is any suggestion you would like to share.

  Regards,

  Hernani

  0
• 

  Philip Taffet

  • April 27, 2021 16:18
  • Edited

  

  

  

  The mathematical description seems to encode \( |V|*\sum_u C_u \) constraints, but it looks like you are only creating \( |V| \) constraints. Could that be the problem?

  0

Please sign in to leave a comment.