Adding conditional constraints to the model

Answered

Qiru MA

May 03, 2023 05:16

• 
• 
• 

I have trouble with implementing the conditional constraint:

# when E_i[i] == m, we want to let G_i_m[i, m] == 1
# when E_i[i] != m, we want to let G_i_m[i, m] == 0

I got this error: gurobipy.GurobiError: Constraint comparison not implemented.

Is there any way to do so? Thanks for your help.

M = range(8) 
N = range(9) 
E_i = model.addVars(len(N), vtype=gp.GRB.INTEGER, name="E_i")
G_i_m = model.addVars(len(N), len(M), vtype=gp.GRB.BINARY, name="G_i_m")

# when E_i[i] == m, we want to let G_i_m[i, m] == 1
# when E_i[i] != m, we want to let G_i_m[i, m] == 0
for i in N:
for m in M:
model.addConstr((E_i[i] - m == 0) >> G_i_m[i, m] >= 1, name="c0_0")
model.addConstr((E_i[i] - m >= 1) >> G_i_m[i, m] == 0, name="c0_1")
model.addConstr((E_i[i] - m <= 1) >> G_i_m[i, m] == 0, name="c0_2")

0

• 

  Jonasz Staszek

  • Community Moderator

  May 03, 2023 06:48

  • 
  • 
  • 

  Hi Qiru,

  as discussed here, indicator constraints require binary variables as indicator variables.

  You may need to add an auxiliary binary variable to represent the relationship between \( E_i \) and \(m\), and then use that binary variable in your indicator constraints.

  If you need more help / example, please let us know.

  Best regards
  Jonasz

  0

Please sign in to leave a comment.