Can gurobi optimize an external function?

Awaiting user input

Eihab Elmukashfi

April 05, 2023 18:43

• 
• 

Hi, I want to use gurobi to optimize the distributed energy resource (DER) capacity to minimize loss. I I use pandapower function to calculate the power loss,  I tried to write a model that recive the DER capacity from gurobi, solve the function via panapower and return the objective value to gurobi again. Is it possible? I elaborated my problem in the following example.

import pandapower as pp
import pandapower.networks as nw
from gurobipy import Model, GRB, quicksum

def loss_function(x):
    net = nw.case33bw()
    pp.create_sgen(net,bus=10,p_mw=x)
    pp.runpp(net)
    return net.res_line["pl_mw"].sum()

 

m=Model()

Gen_capacity = m.addVar(vtype=GRB.CONTINUOUS)
loss = m.addVar(vtype=GRB.CONTINUOUS)

m.update() 

m.setObjective(loss,GRB.MINIMIZE)

m.addConstrs(loss=loss_function(Gen_capacity))

 

0

• 

  Eli Towle

  • Gurobi Staff

  April 05, 2023 19:06

  Unfortunately, Gurobi does not optimize arbitrary external functions like this. Gurobi can solve problems of the form described in the What types of models can Gurobi solve? article. Are you able to explicitly write/reformulate your problem in that form (i.e., with a linear or quadratic objective function and linear, quadratic, and/or second-order cone constraints)?

  0

Please sign in to leave a comment.