Bug Report: MVar with multiple leading 1's in the shape does not accept lb, ub or obj

Florian Peterssen

• February 12, 2021 06:13
• Edited

Currently, it is not possible to add an MVar whose shape has multiple leading 1s with a lower bound of the same shape:

import gurobipy as gp
import numpy as np

shape = (1, 1, 2)
m = gp.Model()
mv1 = m.addMVar(shape, lb=np.ones(shape))

The last line fails with

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "src\gurobipy\model.pxi", line 2903, in gurobipy.Model.addMVar
  File "src\gurobipy\model.pxi", line 2793, in gurobipy.Model.addVars
  File "src\gurobipy\model.pxi", line 280, in gurobipy.__listify.listify
  File "src\gurobipy\model.pxi", line 266, in gurobipy.__listify.flatten
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

It is possible to flatten the array, however, I think this is pretty unintuitive.

mv1 = m.addMVar(shape, lb=np.ones(shape).flatten())

The same goes for the upper bound and the objective coefficients.

I am using Python 3.7 and Gurobi 9.1.1 on Windows 10

0

• 

  Jaromił Najman

  Gurobi Staff

  • February 16, 2021 13:46

  Hi Florian,

  Thank you for bringing this up. We added this issue to our future roadmap and it will be hopefully fixed in a future version of Gurobi.

  For now, using \(\texttt{flatten()}\) is a valid workaround.

  Best regards,
  Jaromił

  0

Please sign in to leave a comment.