2. Installation

Start here to confirm supported platforms, install ADMM, and verify that the Python environment is ready for modeling and solving with ADMM.

Recommended order:

  1. Check Supported Platforms to confirm OS and Python support.

  2. Follow Install ADMM Python Library to install the admm package.

  3. Run the verification step before moving on to the User Guide.

2.1. Supported Platforms

The tables below summarize the supported operating systems and Python versions for the documented Python package.

Supported operating systems

Operating systems

Requirements

Windows

Windows 10 or higher

Linux

GLIBC 2.17 or higher

macOS

Apple Silicon, macOS 12.0 or higher
Supported Python versions

Language

Requirements

Python

Python 3.10 or higher

The examples in this documentation assume a standard scientific Python environment with numpy available. If your platform matches the tables above, continue to Install ADMM Python Library.

2.2. Install ADMM Python Library

You can install ADMM from PyPI or from source.

2.2.1. Install from PyPI

pip install admm

To upgrade an existing installation:

pip install --upgrade admm

2.2.2. Install from Source

Prerequisites:

  • Python >= 3.10

  • C++ compiler (GCC, Clang, or MSVC)

  • admmlib >= 2026.4.9 (admm C++ core dependency library)

Install steps:

git clone https://github.com/alibaba-damo-academy/admm.git
cd admm
pip install . -r requirements.txt

2.2.3. Isolated Python Environment

Using a virtual environment helps avoid dependency conflicts and keeps the installation reproducible. If you use conda, create and activate an environment first, then install the package with pip:

conda create --name admm-py310 python=3.10
conda activate admm-py310
pip install admm

On macOS, make sure the Python environment matches the machine architecture. When using conda, it is worth checking that the reported architecture matches uname -m.

2.2.4. Installation Verification

After installation, first run a short import check:

python -c "import admm; print('ADMM import OK')"

If the command prints ADMM import OK without errors, the Python package is ready to use.

Then run a minimal solve smoke test:

import admm
import numpy as np

model = admm.Model()
x = admm.Var("x", 2)
model.setObjective(admm.sum(admm.square(x - np.array([1.0, 2.0]))))
model.optimize()

print(model.StatusString)
print(x.X)

If this runs without errors and returns a valid solution, the installation is ready for the User Guide.

2.2.5. Troubleshooting

If installation or import fails, check the following points:

  • Confirm that you are using a supported Python version.

  • Make sure the active environment is the one where admm was installed.

  • Upgrade pip before retrying the installation.

  • On macOS, check for architecture mismatches between Python, conda, and the host machine.