Multi-index variable in gurobipy-pandas

Answered

Koen Timmermans

• April 12, 2023 13:15
• Edited

Hi!

I am trying to define a binary decision variable \(x_{ijk}\) using gurobipy-pandas. I am working with two Pandas DataFrames: 1.) task_list, and 2.) vehicle_data. \(x_{ijk}\) is equal to 1 if vehicle \(k\) is driving from task \(i\) to task \(j\). Indices \(i\) and \(j\) therefore result from "task_list" and index \(k\) from "vehicle_data". \(i\) and \(j\) have the same range but \(k\) is much smaller.

To define x in Python, I now have the following code in place:

x = gppd.add_vars(model, ???, vtype=GRB.BINARY, name="x")

Unfortunately I don't know how to define "pandas_obj" on the position of the "???" in the line of code above. I found an example using \(\texttt{pd.MultiIndex( )}\)  online, but I am not sure whether that is the way to go.

Hopefully someone can help me out here. Thank you in advance!

Best regards,

Koen Timmermans

Related to

• gurobipy-pandas

0

• 

  Riley Clement

  Gurobi Staff

  • April 13, 2023 03:36

  Hi Koen,
  
  The second argument to add_vars is the index of the variables.  We want the variable to be indexed by i,j,k so we will want a multi-index.  Let's say A is the set of tasks, and K is the set of vehicles.  You can create the required multi-index using pandas.MultiIndex.from_product

  import gurobipy as gp
  from gurobipy import GRB
  import gurobipy_pandas as gppd
  import numpy as np
  import pandas as pd
  
  A = (1,2,3)
  K = (1,2)
  
  model = gp.Model()
  
  x = gppd.add_vars(model, pd.MultiIndex.from_product((A,A,K), names=("i", "j", "k")), vtype=GRB.BINARY, name="x")

  
  x is then a pandas.Series of gurobi variables:

  j  i  k
  1  1  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     2  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     3  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
  2  1  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     2  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     3  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
  3  1  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     2  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
     3  1    <gurobi.Var *Awaiting Model Update*>
        2    <gurobi.Var *Awaiting Model Update*>
  Name: x, dtype: object

  - Riley

  0
• 

  Koen Timmermans

  • April 13, 2023 06:48

  

  

  

  Hi Riley,

  Thank you for the quick and elaborate response!

  To get to the set A and K that you mention in your example, the index columns of my Pandas DataFrames (task_list and vehicle_data) can be used right?

  import pandas as pd
  task_list = pd.DataFrame()
  vehicle_data = pd.DataFrame()
  
  A = task_list.index.to_list()
  K = vehicle_data.index.to_list()

  Other than that, it is clear to me now. Thanks a lot!

  Best regards,

  Koen Timmermans

  0
• 

  Riley Clement

  Gurobi Staff

  • April 13, 2023 07:20

  Hi Koen,

  No need for the .to_list() conversion, but yes that should work.

  - Riley

  0

Please sign in to leave a comment.