Parametric optimization with Gurobi in python?

Answered

Samuel Mallick

April 17, 2024 12:39

• 
• 
• 

Hello,

Does gurobipy support parametric programming? For example I have an optimization problem with a variable 'x' and a constraint 'x - a <= 0'. The parameter 'a' may appear in many other constraints and in the objective function.

I want to solve this optimization problem many times with different values of 'a'. Is this possible with gurobipy?

I am looking for something similar to what can be done with cvxpy: x = cp.Variable((1,1)), a = cp.Parameter((1,1)).

Thank you,

Samuel

Related to

• Modify

0

• 

  Marika Karbstein

  • Gurobi Staff

  April 18, 2024 07:28 Edited

  Hi Samuel,

  There is no automatic way in gurobipy. But you can update the model by

  • changing the attribute values, for example, RHS,  Obj,  LB, or UB 
  •  using function Model.chgCoeff() to change a coefficient in a constraint

  For your example x - a <= 0, the part could look like

  a=1 # start value
  con = m.addConstr(x - a <= 0)
  m.optimize()
  # change a = 5
  con.RHS=5
  m.optimize()

  Best regards,
  Marika

  0

Please sign in to leave a comment.