Python wheels for NLopt, a library for nonlinear optimization. NLopt provides a common interface to a large collection of optimization algorithms, both global and local, gradient-based and derivative-free, for constrained or unconstrained problems.

This project exists to make installing NLopt in Python as simple as pip install nlopt — no compiler, SWIG, or system NLopt installation required. Prebuilt wheels are published for Windows, macOS, and Linux.

pip install nlopt

Prebuilt wheels are available for Python 3.10+ on Windows, macOS, and Linux (x86_64 and arm64). If no matching wheel is found for your platform, pip will fall back to building from source, which requires SWIG and a C++ compiler.

import nlopt
import numpy as np

def objective(x, grad):
    return x[0] ** 2 + x[1] ** 2

opt = nlopt.opt(nlopt.LN_COBYLA, 2)
opt.set_min_objective(objective)
opt.set_lower_bounds([-10, -10])
opt.set_upper_bounds([10, 10])
opt.set_xtol_rel(1e-6)

x = opt.optimize(np.array([1.0, 1.0]))
print(f"Minimum found at {x}, value = {opt.last_optimum_value()}")

Full API documentation, including the list of supported algorithms, is available at nlopt.readthedocs.io.

Builds run nightly and publish dev versions to TestPyPI so regressions surface before a real release. Stable releases are cut from tagged GitHub Releases and published to PyPI. The vendored extern/nlopt submodule is checked daily against upstream and bumped automatically via PR when a new NLopt version is released.

This project wraps NLopt, whose underlying routines are covered by a mix of licenses (mainly MIT and LGPL) depending on the algorithm — see LICENSE for details. The Python packaging in this repository is licensed under MIT.