Optyx

Optimization that reads like Python.

📚 Documentation · 🚀 Quickstart · 💡 Examples

from optyx import Variable, Problem

x = Variable("x", lb=0)
y = Variable("y", lb=0)

solution = (
    Problem()
    .minimize(x**2 + y**2)
    .subject_to(x + y >= 1)
    .solve()
)
# x=0.5, y=0.5

from scipy.optimize import minimize
import numpy as np

def objective(v):
    return v[0]**2 + v[1]**2

def gradient(v):  # manual!
    return np.array([2*v[0], 2*v[1]])

result = minimize(
    objective, x0=[1, 1], jac=gradient,
    method='SLSQP',
    bounds=[(0, None), (0, None)],
    constraints={'type': 'ineq',
                 'fun': lambda v: v[0]+v[1]-1}
)

Your optimization code should read like your math. With Optyx, x + y >= 1 is exactly that—not a lambda buried in a constraint dictionary.

Why Optyx?

Python has excellent optimization libraries. SciPy provides algorithms. CVXPY handles convex problems. Pyomo scales to industrial applications.

Optyx takes a different path: radical simplicity.

• Write problems as you think them — x**2 + y**2 not lambda v: v[0]**2 + v[1]**2
• Never compute gradients by hand — symbolic autodiff handles derivatives
• Skip solver configuration — sensible defaults, automatic solver selection

Being Honest

Optyx is young and opinionated. It's not a replacement for specialized tools:

Optyx does support MILP (via HiGHS), sparse LPs with 100k+ variables, and solver callbacks—but if you need industrial-grade MIP with cutting planes, a dedicated solver is the right choice.

Installation

pip install optyx

Requires Python 3.12+, NumPy ≥2.0, SciPy ≥1.7.

Quick Examples

Constrained Quadratic

from optyx import Variable, Problem

x = Variable("x", lb=0)
y = Variable("y", lb=0)

solution = (
    Problem()
    .minimize(x**2 + y**2)
    .subject_to(x + y >= 1)
    .solve()
)
# x=0.5, y=0.5, objective=0.5

Portfolio Optimization

from optyx import Variable, Problem

# Asset weights
tech = Variable("tech", lb=0, ub=1)
energy = Variable("energy", lb=0, ub=1)
finance = Variable("finance", lb=0, ub=1)

# Expected returns and risk (simplified)
returns = 0.12*tech + 0.08*energy + 0.10*finance
risk = tech**2 + energy**2 + finance**2  # variance proxy

solution = (
    Problem()
    .minimize(risk)
    .subject_to(returns >= 0.09)              # minimum return
    .subject_to((tech + energy + finance).eq(1))  # fully invested
    .solve()
)

Autodiff Just Works

from optyx import Variable
from optyx.core.autodiff import gradient

x = Variable("x")
f = x**3 + 2*x**2 - 5*x + 3

df = gradient(f, x)  # Symbolic: 3x² + 4x - 5
print(df.evaluate({"x": 2.0}))  # 15.0

Mixed-Integer Programming

from optyx import BinaryVariable, Problem

# Decision: pick an item (1) or leave it out (0)
x1 = BinaryVariable("x1")
x2 = BinaryVariable("x2")
x3 = BinaryVariable("x3")
x4 = BinaryVariable("x4")
x5 = BinaryVariable("x5")

value = 10*x1 + 20*x2 + 15*x3 + 25*x4 + 30*x5
weight = 5*x1 + 10*x2 + 8*x3 + 12*x4 + 15*x5

solution = (
    Problem()
    .maximize(value)
    .subject_to(weight <= 30)
    .solve()
)
# Automatically routes to a MILP solver

Features at a Glance

See the documentation for the full API reference, tutorials, and real-world examples.

What's Next

Optyx is actively evolving:

• MIQP / MINLP support — Quadratic and nonlinear MIP via native HiGHS or Gurobi
• MPS format I/O — Import and export MPS files for solver interop
• More solvers — IPOPT integration for large-scale NLP
• Better debugging — Infeasibility diagnostics and model inspection

See the roadmap for details.

Contributing

git clone https://github.com/optyx-dev/optyx.git
cd optyx
uv sync
uv run pytest

Contributions welcome! See our contributing guide.

License

MIT