Why are binary Variables getting float results. Is there any command to prevent it?

Answered

BISWAJIT KAR

October 08, 2023 17:03

• 
• 
• 

Hello. 

I have defined two types of variables in my model running in different iterations. To restrict the variable values, I have considered both ways like binary and Integer variables with lb=0 and ub=1. 

However, for some iterations, the variables are taking float values. Is there any way to prevent it? Please Suggest.

 

Related to

• Solutions

0

• 

  Jaromił Najman

  • Gurobi Staff

  October 09, 2023 09:33

  Hi Biswajit,

  It is possible, that due to machine precision and numerical inaccuracies caused by, e.g., so-called trickle-flows, that discrete variables take float values in an optimal solution. To avoid this you could either decrease the IntFeasTol parameter value or use the IntegralityFocus parameter.

  I recommend using (and reading the description of) the IntegralityFocus parameter even tho it might have a visible impact on performance

  Best regards, 
  Jaromił

  0
• 

  Gwyneth Butera

  • Gurobi Staff

  October 10, 2023 11:44

  In addition to Jaromil's reply, please refer to Why does Gurobi sometimes return non-integral values for integer variables?

  0

Please sign in to leave a comment.