Gurobi: Invalid data in vars array

回答済み

Shahrzad Valizadeh

• 2024年05月16日 13:17

I have a vector of 100 demand values, for each of them i want to define specific constraints. Here is how I've defined my variables.

        b1 = {}
        for r in blood_types:
            for m in M:
                for demand_index in range(100):
                    b1[r, "1", m, demand_index] = Issuing_model.addVar(vtype=GRB.INTEGER, name=f'b1_{r}_{"1"}_{m}_{demand_index}')

        I1 = {}
        for r in blood_types:
            for m in m1:
                I1[r, "1", m] = Issuing_model.addVar(vtype=GRB.INTEGER, name=f'I1_{r}_{"1"}_{m}')

while:

M = list(range(41, 0, -1))

m1 = list(range(42, 0, -1))

But the following constraint, gives me the error of: 

GurobiError: Invalid data in vars array

        for demand_index, demand in enumerate(next_demands_OPM):
            next_demand_OPM = demand
            for m in M:
                if m == 1:
                    Issuing_model.addConstr(b1["O+", "1", m, demand_index] == gp.max_(0, I1["O+", "1", m+1] - next_demand_OPM))
                else:
                    Issuing_model.addConstr(b1["O+", "1", m, demand_index] == gp.max_(0, I1["O+", "1", m+1] - (next_demand_OPM - sum(I1["O+", "1", j] for j in range(2, m+1)))))

Does anyone have any feedback?

関連するエンティティ

• Python
• gurobipy-pandas

0

• 

  Marika Karbstein

  Gurobi Staff

  • 2024年05月16日 14:08

  It looks like you want to combine a max-constraint with a linear expression. This needs to be done in two steps and you need an additional auxiliary variable, e.g.

  Issuing_model.addConstr(aux["O+", "1", m, demand_index] == gp.max_(0, I1["O+", "1", m+1]))
  Issuing_model.addConstr(b1["O+", "1", m, demand_index] == aux["O+", "1", m, demand_index] - (11 - sum(I1["O+", "1", j] for j in range(2, m+1))))

  0
• 

  Shahrzad Valizadeh

  • 2024年05月17日 00:10

  

  

  Thank you Marika for your time. In fact, I'm trying to write the constraint provided in the inserted image. I tried your suggestion and defined auxiliary variable to the model, but the problem stays the same.

      b1_{O+, 1, m} = (I1_{O+, 1, m+1} - (next demand OPM - \sum_{j=2}^{m+1}(I1_{O+, 1, j}))^{+})^{+}
  
  

  0
• 

  Marika Karbstein

  Gurobi Staff

  • 2024年05月17日 05:55

  Please provide a minimal reproducible example, see Tutorial: Preparing a Minimal Reproducible Example

  0
• 

  Shahrzad Valizadeh

  • 2024年05月17日 06:04

  

  

  import sys
  import gurobipy as gp
  from gurobipy import Model, GRB
  from datetime import datetime
  from gurobipy import quicksum
  from gurobipy import max_
  from gurobipy import LinExpr 

  next_demands_OPM is a vector of 100 demand values.      

  m1 = list(range(42, 0, -1))

  M = list(range(41, 0, -1))

  z11 = {}
          for r in blood_types:
              for m in M:
                  for demand_index in range(100):
                      z11[r, "1", m, demand_index] = Issuing_model.addVar(lb=-GRB.INFINITY, ub=GRB.INFINITY, name=f'z11_{r}_{"1"}_{m}_{demand_index}')
  z12 = {}
          for r in blood_types:
              for m in M:
                  for demand_index in range(100):
                      z12[r, "1", m, demand_index] = Issuing_model.addVar(lb=-GRB.INFINITY, ub=GRB.INFINITY, name=f'z12_{r}_{"1"}_{m}_{demand_index}')
  z13 = {}
          for r in blood_types:
              for m in M:
                  for demand_index in range(100):
                      z13[r, "1", m, demand_index] = Issuing_model.addVar(lb=-GRB.INFINITY, ub=GRB.INFINITY, name=f'z13_{r}_{"1"}_{m}_{demand_index}')                    
  z14 = {}
          for r in blood_types:
              for m in M:
                  for demand_index in range(100):
                      z14[r, "1", m, demand_index] = Issuing_model.addVar(lb=-GRB.INFINITY, ub=GRB.INFINITY, name=f'z14_{r}_{"1"}_{m}_{demand_index}')         

  b1 = {}
          for r in blood_types:
              for m in M:
                  for demand_index in range(100):
                      b1[r, "1", m, demand_index] = Issuing_model.addVar(vtype=GRB.INTEGER, name=f'b1_{r}_{"1"}_{m}_{demand_index}')

  I1 = {}
          for r in blood_types:
              for m in m1:
                  I1[r, "1", m] = Issuing_model.addVar(vtype=GRB.INTEGER, name=f'I1_{r}_{"1"}_{m}')     

       for demand_index, demand in enumerate(next_demands_OPM):
              next_demand_OPM = demand
              for m in M:
                  if m == 1:
                      Issuing_model.addConstr(z14["O+", "1", m, demand_index] == gp.max_(0, I1["O+", "1", m+1] - next_demand_OPM))
                      Issuing_model.addConstr(b1["O+", "1", m, demand_index]  == z14["O+", "1", m, demand_index])
                  else:
                      Issuing_model.addConstr(z11["O+", "1", m, demand_index] == next_demand_OPM - gp.quicksum(I1["O+", "1", j] for j in range(2, m+1)))
                      Issuing_model.addConstr(z12["O+", "1", m, demand_index] == gp.max_(0, z11["O+", "1", m, demand_index]))
                      Issuing_model.addConstr(z13["O+", "1", m, demand_index] == I1["O+", "1", m+1] - z12["O+", "1", m, demand_index])
                      Issuing_model.addConstr(b1["O+", "1", m, demand_index]  == gp.max_(0, z13["O+", "1", m, demand_index]))

  0
• 

  Marika Karbstein

  Gurobi Staff

  • 2024年05月17日 06:35

  Please check how Model.addGenConstrMax() is used.

  gp.max_(0, I1["O+", "1", m+1] - next_demand_OPM))

  is not possible. The input has to be variables over which the MAX will be taken.
  I1["O+", "1", m+1] - next_demand_OPM) is a linear expression and needs to be assigned to an (auxiliary) variable first.
  So, you probably want to have

  Issuing_model.addConstr(z14["O+", "1", m, demand_index] == I1["O+", "1", m+1] - next_demand_OPM)
  Issuing_model.addConstr(b1["O+", "1", m, demand_index]  == gp.max_(0, z14["O+", "1", m, demand_index]))

  0
• 

  Shahrzad Valizadeh

  • 2024年05月17日 07:08

  

  

  Thank you so much Marika for your time and suggestions. That was so valuable to me.

  0

サインインしてコメントを残してください。