Formulating MIP with FICO xpress using Pyomo

Question

How do I solve an MIP model with FICO Xpress, via Pyomo?

In addition, I want to set parameters of Xpress like miptol and miprelstop. How do I do that in Pyomo?

Answer 1

In order to mark a variable as binary, you have to define it to be within the Binary constraint:

model.z = pyo.Var(bounds=(0, 1)), within=pyo.Binary)

Setting controls is done with the options attribute of the solver. Here is a short sample that creates a simple model, sets options, and solves.

import pyomo.environ as pyo

model = pyo.ConcreteModel()
model.x = pyo.Var(bounds=(0, 1))
model.y = pyo.Var(bounds=(0, 1))
model.z = pyo.Var(bounds=(0, 1), within=pyo.Binary)
model.c = pyo.Constraint(expr = model.z >= 0.5)
model.obj = pyo.Objective(expr = model.x - 2 * model.y + 3 * model.z,
                          sense = pyo.minimize)

opt = pyo.SolverFactory('xpress_direct')

opt.options['miprelstop'] = 1e-9
opt.options['miptol'] = 0.5

results = opt.solve(model)

print('x =', pyo.value(model.x))
print('y =', pyo.value(model.y))
print('z =', pyo.value(model.z))

This outputs (among other things)

x = 0.0
y = 1.0
z = 1.0

which shows that the binary condition on z was respected (otherwise, the optimal value for z would have been 0.5). It also shows the modified parameter values in the output log.