How to use addGenConstrMax function?

Answered

Menglei Ji

October 30, 2023 14:47

• 
• 
• 

Hi there,

when I use this sentence to add the infinity norm constraint:

model.addGenConstrMax(b, [a_pos[i] + a_neg[i] for i in I])   

there is an error: 'GurobiError: Invalid data in vars array'. Could you please tell me how to correct this?

I also try to use model.addGenConstrNorm, but it also shows the same error.

 

Thanks,

Menglei

   

Related to

• General Constraints

0

• 

  Eli Towle

  • Gurobi Staff

  October 30, 2023 15:03

  Model.addGenConstrMax() is used to set one variable equal to the maximum of a set of variables, and possibly a constant. The \(\texttt{a_pos[i] + a_neg[i]}\) terms are (likely linear) expressions, not variables.

  To resolve the issue, introduce auxiliary variables for each of the \(\texttt{a_pos[i] + a_neg[i]}\) terms, then use those auxiliary variables in your call to Model.addGenConstrMax(). For example:

  # Add lower bound of -GRB.INFINITY if a_pos[i] + a_neg[i] can be negative
  u = model.addVars(I, name="u")
  
  model.addConstrs((a_pos[i] + a_neg[i] == u[i] for i in I), name="define_u")
  model.addGenConstrMax(b, u, name="define_b")

  0
• 

  Menglei Ji

  October 31, 2023 02:52

  • 
  • 
  • 

  Hi Eli,

  Thanks for your response. I still have a question about addGenConstrNorm.

        alpha = model.addVars (I, S,lb=-GRB.INFINITY, vtype=GRB.CONTINUOUS, name= 'alpha')
        beta = model.addVars (S, lb=0,vtype=GRB.CONTINUOUS, name= 'beta')

        self.model_dro.addGenConstrNorm(beta, alpha, Q, 'normconstr') # Q can be 1,2, GRB.INFINITY.

  The error is:

  TypeError: {1: <gurobi.Var *Awaiting Model Update*>, 2: <gurobi.Var *Awaiting Model Update*>, 3: <gurobi.Var *Awaiting Model Update*>, 4: <gurobi.Var *Awaiting Model Update*>} is not a variable.

  I am so confused about the command. Thanks for your help!

   

  Best,

  Menglei

  0
• 

  Eli Towle

  • Gurobi Staff

  October 31, 2023 15:46

  The first argument to Model.addGenConstrNorm() should be a single Var object, not a tupledict. If you are trying to add a norm constraint for every \( s \in S \), you will have to define the norm constraint for each individual Var object in \(\texttt{beta}\). For example, using the norm() general constraint helper function:

  import gurobipy as gp
  
  self.model_dro.addConstrs(
      (beta[s] == gp.norm(alpha.select("*", s), Q) for s in S), "normconstr"
  )

  Also, your code adds the variables to Model object \(\texttt{model}\), then adds constraints to Model object \(\texttt{self.model_dro}\). Maybe they are the same model, but I will mention it anyways in case you run into a "Variable not in model" error.

  0

Please sign in to leave a comment.