gethighs

A Python interface to use HiGHS executable files with Pyomo.

Current alternatives

Appsi solver

Pyomo has currently HiGHS appsi solver implemented, which works very well for systems in which highspy is available (and stable).

You can install highspy directly from pip (make sure pyomo and pybind11 are installed)

pip install highspy

And import the appsi solver from the following path:

from pyomo.contrib.appsi.solvers import Highs

Use it as any pyomo solver from solver factory and be happy! Unless you have some trouble with highspy... And you might, as I did.

If that's your case, keep with me a few more lines.

Downloading executables

First, make sure you have a highs executable file available in your computer.

You can find HiGHS static executable files pre-compiled for several systems on this link.

More information is available on HiGHS github page.

After dowloading the executables, it helps if they are included in system (or user) path. You might need no reinitialize your computer after this.

Installing gethighs

gethighs is a small Python package which I developed for using HiGHS directly from executable files via the Pyomo interface. You can install it via pip + the github link:

pip install git+https://github.com/bruscalia/gethighs

And also use it almost as a native pyomo solver.

Usage

You can see more details in the example notebook.

Consider this simple example:

$$ \begin{align} \text{maximize}~~ & 5 x_{1} + 4 x_{2} \\ \text{subject to}~~ & 2 x_{1} + 3 x_{2} \leq 12 \\ & 2 x_{1} + x_{2} \leq 6 \\ & x_{i} \geq 0 & \forall ~ i \in { 1, 2 } \\ & x_{i} \in \mathbb{Z} & \forall ~ i \in { 1, 2 } \end{align} $$

Import the required libraries:

import pyomo.environ as pyo
from gethighs import HiGHS

Instantiate the model.

# Initialize model
model = pyo.ConcreteModel()

# Set
model.I = pyo.Set(initialize=[1, 2])

# Variables
model.x = pyo.Var(model.I, within=pyo.NonNegativeIntegers)

# Constraints
model.c1 = pyo.Constraint(expr=2 * model.x[1] + 3 * model.x[2] <= 12)
model.c2 = pyo.Constraint(expr=2 * model.x[1] + model.x[2] <= 6)

# Objective
model.obj = pyo.Objective(expr=5 * model.x[1] + 4 * model.x[2], sense=pyo.maximize)

Instantiate and configure the solver.

solver = HiGHS(time_limit=10, mip_heuristic_effort=0.2, mip_detect_symmetry="on")

Solve the model.

results = solver.solve(model)
print(results)

status: Optimal
primal_solutions: Feasible
objective: 18.0

model.x.display()

    x : Size=2, Index=I
    Key : Lower : Value : Upper : Fixed : Stale : Domain
      1 :     0 :   2.0 :  None : False : False : NonNegativeIntegers
      2 :     0 :   2.0 :  None : False : False : NonNegativeIntegers

Contact

You can reach out to me at: [email protected]