constraint for inequality

• 

  Alex K  (Report)

  November 26, 2020 23:30

  Hi Le Hoa,

  What does your  myfunc_ look like?

  My approach would be to reframe your constraint as

  # Collect the non-variable terms
  z = (a - b - c)/d + 1
  
  # Reframe the constraint
  z  >= x

   

  In this form it's a bit easier to see that if z >= 1 the constraint will always be satisfied (since the maximum value of x is 1).  So you could do this: 

  # Function to evaluate z >= x (where z = (a - b - c)/d + 1)
  def myfunc_(z):
      if z < 1:
          return lambda x: x == 0
      else:
          return lambda x: True
  
  const = dwavebinarycsp.Constraint.from_func(myfunc_(z), ['x'], dwavebinarycsp.BINARY, name='myfunc_{}'.format(z))

   

  I think it's worth mentioning that there's a more performant way of defining this constraint as well. When the constraint is always satisfied we don't actually need the constraint at all. So you could evaluate z while you're constructing your BQM. If z >= 1 you can fix the variable to true using fix_variable(). 

  Let me know if this helps and if you have any more questions! 

  0

  Comment actions Permalink

Please sign in to leave a comment.