Skip to main content

Value Update after every iteration

Awaiting user input
Param Patel
First Comment
First Question

Hello Community,

I am facing issues while updating the values of energy variable e1t after every iteration. Please refer to the following code and output, 

for i in range(len(v1_t)):
        for j in range(len(tk)):
            res = model.addVar(lb=0, ub=50000, vtype=gb.GRB.INTEGER, name=f'volts*current_{i}_{j}')
            final_res = model.addVar(lb=0, ub=50, vtype=gb.GRB.INTEGER, name=f'round_result_{i}_{j}')
            model.addConstr(res == r1t[i, j] * v * delta_t)
            model.addConstr(final_res == res / 1000)
            # Constraint to update pending energy based on cumulative charging completed till that step
            if tk[i]<=tk[j]<=T_full_charge[i]:
                if j > 0:
                    model.addConstr(e1t[i, j] == e1[j-1] - final_res)
                else:
                    model.addConstr(e1t[i, j] == e1[i] - final_res)
            else:
                model.addConstr(e1t[i,j]==0)
 
Output:e1t_0_0 = 50.0, e1t_0_1 = 50.0
0

Comments

7 comments

Date Votes
  • Jaromił Najman
    Gurobi Staff Gurobi Staff
    • Edited

    Hi Param,

    Could you please clarify what exactly the issue is? Is it about not getting the correct solution point? Do you get an error during model building?

    Note that you can always write your model to a human-readable LP file via the write method

    model.write("myModel.lp")

    You can open the file \(\texttt{myModel.lp}\) in any standard text editor and analyze whether the model looks as expected.

    Maybe the Knowledge Base article How do I diagnose a wrong solution? might be helpful.

    Moreover please note that your code snippet is not a minimal reproducible example.

    Best regards, 
    Jaromił

    0
  • Param Patel
    First Comment
    First Question

    Thank you for your reply Jaromil.

    The problem is with formulating the variables. Here,variable r[i,j] is not being assigned properly. Thus, the model always picks variable r as 0 at all timesteps.

    def model_variables_and_constraints(model, tk, v1_t, v2_t, e1, e2, v, T_full_charge, T_top_up, delta_t):
        r1t = {}
        e1t = {}
        # current_set = [8, 16, 24, 32, 48, 64]
        b1 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_8a')
        b2 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_16a')
        b3 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_24a')
        b4 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_32a')
        b5 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_48a')
        b6 = model.addVar(vtype=gb.GRB.BINARY, name='binary_var_current_64a')
     
        model.addConstr(gb.quicksum([b1,b2,b3,b4,b5,b6])==1)
        # Add the charging variables r and e for full charge and top-up EV
        for i in range(len(v1_t)):
            for j in range(len(tk)):
                r1t[i, j] = model.addVar(lb=0, ub=64, vtype=gb.GRB.INTEGER, name=f'EV_{i}_current_{j}')
                model.addConstr(r1t[i,j] == gb.quicksum([8*b1,16*b2,24*b3,32*b4,48*b5,64*b6]))
                e1t[i, j] = model.addVar(lb=0, ub=e1[i], vtype=gb.GRB.INTEGER, name=f'ChargingVariable_e1_{i}_{j}')
                if tk[i] <= tk[j] <= T_full_charge[i]:
                    model.addConstr(r1t[i, j] == 0)
     
        # Define constraints
        for i in range(len(v1_t)):
            for j in range(len(tk)):
                res = model.addVar(lb=0, ub=50000, vtype=gb.GRB.INTEGER, name=f'volts*current_{i}_{j}')
                final_res = model.addVar(lb=0, ub=50, vtype=gb.GRB.INTEGER, name=f'round_result_{i}_{j}')
                model.addConstr(res == r1t[i, j] * v * delta_t)
                model.addConstr(final_res == res / 1000)
                # Constraint to update pending energy based on cumulative charging completed till that step
                if tk[i]<=tk[j]<=T_full_charge[i]:
                    if j > 0:
                        model.addConstr(e1t[i, j] == e1[j-1] - final_res)
                    else:
                        model.addConstr(e1t[i, j] == e1[i] - final_res)
                else:
                    model.addConstr(e1t[i,j]==0)
      return r1t, e1t
    0
  • Jaromił Najman
    Gurobi Staff Gurobi Staff

    I do not see any variable \(\texttt{r[i,j]}\) in your code snippet. Do you mean \(\texttt{r1t[i,j]}\)?

    0
  • Param Patel
    First Comment
    First Question

    Sorry, yes I meant r1t[i,j]

    0
  • Jaromił Najman
    Gurobi Staff Gurobi Staff

    Hi Param,

    The constraints

    model.addConstr(gb.quicksum([b1,b2,b3,b4,b5,b6])==1)
    ...
    model.addConstr(r1t[i,j] == gb.quicksum([8*b1,16*b2,24*b3,32*b4,48*b5,64*b6]))

    should make sure that each \(\texttt{r1t}\) variable has at least value \(\texttt{8}\).

    Could you please write your model to an LP file and analyze whether it looks as expected? You can do so via

    model.write("myModel.lp")

    You can open the file \(\texttt{myModel.lp}\) in any standard text editor.

    Please note that the constraints

    if tk[i] <= tk[j] <= T_full_charge[i]:
      model.addConstr(r1t[i, j] == 0)

    are contradicting the previous constraint.

    Best regards, 
    Jaromił

    0
  • Param Patel
    First Comment
    First Question

    Thankyou for your reply Jaromil. I changed the constraint which was incorrect. But the issue still persists. The thing is I also tried the same problem using PuLP solver, I can find an optimal solution using that. Also, when I write the model, there is a problem in variable initialization. You can see it here.

    Maximize
      - 119.5 EV_0_current_0 - 0.5 ChargingVariable_e1_0_0
       - 119.5 EV_0_current_1 - 0.5 ChargingVariable_e1_0_1
       - 119.5 EV_0_current_2 - 0.5 ChargingVariable_e1_0_2
       - 120 EV_1_current_0 - 0.5 ChargingVariable_e1_1_0 - 120 EV_1_current_1
       - 0.5 ChargingVariable_e1_1_1 - 120 EV_1_current_2
       - 0.5 ChargingVariable_e1_1_2 - 120 EV_2_current_0
       - 0.5 ChargingVariable_e1_2_0 - 120 EV_2_current_1
       - 0.5 ChargingVariable_e1_2_1 - 120 EV_2_current_2
       - 0.5 ChargingVariable_e1_2_2 - 119.5 EV_topup_0_current_0
       - 0.5 ChargingVariable_e2_0_0 - 119.5 EV_topup_0_current_1
       - 0.5 ChargingVariable_e2_0_1 - 119.5 EV_topup_0_current_2
       - 0.5 ChargingVariable_e2_0_2 - 120 EV_topup_1_current_0
       - 0.5 ChargingVariable_e2_1_0 - 120 EV_topup_1_current_1
       - 0.5 ChargingVariable_e2_1_1 - 120 EV_topup_1_current_2
       - 0.5 ChargingVariable_e2_1_2 - 120 EV_topup_2_current_0
       - 0.5 ChargingVariable_e2_2_0 - 120 EV_topup_2_current_1
       - 0.5 ChargingVariable_e2_2_1 - 120 EV_topup_2_current_2
       - 0.5 ChargingVariable_e2_2_2
    Subject To
     R0: binary_var_current_8a + binary_var_current_16a
       + binary_var_current_24a + binary_var_current_32a
       + binary_var_current_48a + binary_var_current_64a = 1
     R1: - 8 binary_var_current_8a - 16 binary_var_current_16a
       - 24 binary_var_current_24a - 32 binary_var_current_32a
       - 48 binary_var_current_48a - 64 binary_var_current_64a + EV_0_current_0
       = 0. 

    I modified the constraints now:

     model.addConstr(gb.quicksum([b1,b2,b3,b4,b5,b6])==1)
        # Add the charging variables r and e for full charge and top-up EV
        for i in range(len(v1_t)):
            for j in range(len(tk)):
                r1t[i, j] = model.addVar(lb=0, ub=64, vtype=gb.GRB.INTEGER, name=f'EV_{i}_current_{j}')
                e1t[i, j] = model.addVar(lb=0, ub=e1[i], vtype=gb.GRB.INTEGER, name=f'ChargingVariable_e1_{i}_{j}')
                if tk[i] <= tk[j] <= T_full_charge[i]:
                    model.addConstr(r1t[i,j] == gb.quicksum([8*b1,16*b2,24*b3,32*b4,48*b5,64*b6]))
                else:
                    model.addConstr(r1t[i, j] == 0)

        for i in range(len(v2_t)):
            for j in range(len(tk)):
                r2t[i, j] = model.addVar(lb=0, ub=64, vtype=gb.GRB.INTEGER, name=f'EV_topup_{i}_current_{j}')
                e2t[i, j] = model.addVar(lb=0, ub=e2[i], vtype=gb.GRB.INTEGER, name=f'ChargingVariable_e2_{i}_{j}')
                if tk[i] <= tk[j] <= T_top_up[i]:
                    model.addConstr(r2t[i,j] == gb.quicksum([8*b1,16*b2,24*b3,32*b4,48*b5,64*b6]))
                else:
                    model.addConstr(r2t[i, j] == 0)
    0
  • Jaromił Najman
    Gurobi Staff Gurobi Staff

    The thing is I also tried the same problem using PuLP solver

    Can you write the LP generated by PuLP to an LP file and compare with the one you constructed in gurobipy?

    Alternatively, you could extract the solution point from PuLP and fix all variables in your gurobipy model to see whether the solution point is also feasible in your gurobipy model.

    Also, when I write the model, there is a problem in variable initialization. You can see it here.

    I cannot see any issue. Do you mean that an error occurs? Could you please point me to the issue?

    0

Please sign in to leave a comment.

Didn't find what you were looking for?

New post