How to extract solution values to separate matrices/dataframes in Python?

Answered

Aleksandrs Smilgins

July 15, 2023 14:52

• 
• 
• 

Hi,

I am new to Gurobi and Python. I have a Python code that in the end gives me the solution:

for v in my_model.getVars():
    print(v.varName, v.x)

However, this is a long vector where I have several variables, and they are two-dimensional:

lambda[0,0]

lambda[0,2]

lambda[0,3]

lambda[1,0]

...

lambda[50,3]

group[0,0]

group[0,1]

group[0,2]

group[1,0]

...

group[50,2]

 

Is there a way to save the output in two matrices/dataframes: lambda and group which will then be 2 dimensional (each will be a matrix or dataframe with 50 rows and 3 columns)? In general, there is not necessarily 50 rows and 3 columns - these numbers can vary. Also, there may be other variables than lambda and group. 

Thank you!

Related to

• Pandas

0

• 

  Riley Clement

  • Gurobi Staff

  July 17, 2023 09:44

  Hi Aleksandrs,

  You can use numpy.reshape and Model.getAttr to do this concisely.  For example:

  # setup
  import gurobipy as gp
  import numpy as np
  
  m = gp.Model()
  x = m.addVars(range(5), range(3), vtype="B", obj=np.random.random(size=(5,3)), name="v")
  m.addConstr(gp.quicksum(x.values())==7)
  m.optimize()
  
  # print variable names in numpy matrix to verify the result of reshape
  print(np.reshape(m.getAttr("VarName", m.getVars()), (5,3)))
  
  # print variable values in solution in numpy matrix
  print(np.reshape(m.getAttr("X", m.getVars()), (5,3)))

  The print statements give

  array([['v[0,0]', 'v[0,1]', 'v[0,2]'],
         ['v[1,0]', 'v[1,1]', 'v[1,2]'],
         ['v[2,0]', 'v[2,1]', 'v[2,2]'],
         ['v[3,0]', 'v[3,1]', 'v[3,2]'],
         ['v[4,0]', 'v[4,1]', 'v[4,2]']], dtype='<U6')

  and

  array([[1., 0., 1.],
         [0., 0., 0.],
         [1., 0., 0.],
         [0., 0., 1.],
         [1., 1., 1.]])

  respectively.

  - Riley

   

  2

Please sign in to leave a comment.