Constraint on decision variables in a multi period scenario

Answered

Chaolong Chen

August 22, 2023 09:08 Edited

• 
• 
• 

I have the following parameters:

period = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]
facility = ['f1', 'f2', 'f3', 'f4', 'f5']

I have declared a binary decision variable as follows:

X_f = m.addVars(facility, period, vtype=GRB.BINARY, name="X_f")

In a multi period facility location problem, I have the following constraints:

• for every facility, if X_f have the boolean value of 0 in period 0, it remains 0 for all of the periods
• for every facility, if X_f has the boolean value 1 in period 0, then it remains 1 for all of the periods

Basically, it means that once a facility has been opened in period 0, it should not be closed; and if a facility is closed in period 0, then it should remain closed.

I tried to implement this as follows:

for f in facility:
    if X_f[f, 0].X == 0:
        m.addConstrs((X_f[f, t+1] == 0 for t in range(0, 11)))

    if X_f[f, 0].X == 1:
        m.addConstrs((X_f[f, t+1] == 1 for t in range(0, 11)))

m.optimize()

However, I get the error

AttributeError: Index out of range for attribute 'X'

I understand this is due to the fact that X_f[f, 0].X only has value after the optimization has run.
Is there another way to implement this constraint in a multi period setting?

Related to

• MIP
• Error
• Modeling

0

• 

  Silke Horn

  • Gurobi Staff

  August 22, 2023 10:00

  Hi,

  You are correct that you cannot access the X attribute before running the optimization. Instead, you can use indicator constraints to model something like this.

  This article contains additional information on how to model conditional constraints.

  Silke

   

  0

Please sign in to leave a comment.