Skip to content

Installation

The examples on this site target VAMOS 1.5.0. Use an isolated Python environment and the pinned commands below to follow the same release.

Supported Python and operating systems

Operating system Python versions covered by CI Installed-wheel release checks
Linux 3.11, 3.12, 3.13, 3.14 Core on 3.11 and 3.13; core and compute on 3.12 and 3.14
Windows 3.12, 3.13, 3.14 Core on 3.12
macOS 3.12, 3.13, 3.14 Core on 3.12

These are the combinations in the project's CI and release verification matrices, not a promise for every version of every optional dependency. Other interpreter/platform combinations are unclaimed. See Known limitations.

VAMOS 1.2.0: CI tests Python 3.13 and 3.14 on the three systems, and the release checks install the wheel on both. VAMOS 1.1.0 was tested on Python 3.10, 3.11 and 3.12 only.

VAMOS 1.3.0: VAMOS requires Python 3.11 or newer, as announced with 1.2.0, and CI no longer tests Python 3.10. pip does not install it on Python 3.10, where an unpinned install selects the newest 1.2 release; see Stability and versioning.

Core package

Create an isolated environment and install the package from PyPI:

python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install "vamos-optimization==1.5.0"
py -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install --upgrade pip
python -m pip install "vamos-optimization==1.5.0"

Verify the installation:

python -c "import vamos; print(vamos.__version__)"
vamos check

The version command must print 1.5.0 for this release. To install the latest published version instead, use python -m pip install --upgrade vamos-optimization and consult the documentation for that version. Pinning VAMOS alone does not freeze dependencies; retain the environment record for scientific runs (see Run artifacts and replay).

Optional extras

Install only the capability groups you use:

Extra Purpose Interface status
compute Numba kernels, MooCore indicators, and Dask evaluation Stable backend selection; Experimental Dask integration
research Third-party research baselines and benchmarks Experimental
analysis Data frames, plotting, notebooks, and analysis helpers Experimental helpers
tuning Optional model-based tuning backends Experimental
studio Panel-based local Studio Experimental; see Studio for release limitations
examples Dependencies used by selected examples Depends on the example's interfaces
modact VAMOS 1.5.1: the authors' modact package (GPL-3.0), which evaluates the MODAct problems Registered problems; all leaves it out
docs Documentation toolchain Developer tooling
dev Tests, linting, typing, building, and docs checks Developer tooling
all All optional and development groups Mixed; does not change interface guarantees

VAMOS 1.2.0: the extras install only what VAMOS and its examples use, and no other optimization framework. The research extra (pymoo, jMetalPy, DEAP, Platypus and pygmo) is removed, with the baselines it served. analysis installs pandas, matplotlib, Plotly and scikit-learn. examples adds seaborn, NetworkX, IPython, ipykernel and ipywidgets for the examples and notebooks; in VAMOS 1.1.0 they came with analysis, so a notebook now needs analysis and examples. all installs the user extras (compute, analysis, tuning, studio, examples, openai, anthropic and gemini) and no longer the dev and docs tooling.

VAMOS 1.4.0: the openai, anthropic and gemini extras are removed, with vamos assist and Studio's Problem Builder, the only parts of VAMOS that used them. No extra, all included, installs an LLM client.

VAMOS 1.5.1: the modact extra installs the modact package of Picard and Schiffmann, which evaluates the 20 MODAct problems. It is licensed under the GPL-3.0, unlike VAMOS (MIT), so all does not install it: install the extra, or python -m pip install modact, to run those problems. VAMOS contains none of its code and imports it only when a MODAct problem is created.

For example:

python -m pip install "vamos-optimization[compute,analysis]==1.5.0"

An explicit optional backend fails clearly when its dependency is unavailable. VAMOS does not install packages during optimization, verification, or replay. Installing an extra does not make its APIs stable. The stability policy defines Stable, Experimental, and Internal interfaces.

Install from a source checkout

Only use an editable install when developing VAMOS or modifying repository examples. First obtain the source and enter its root:

git clone https://github.com/vamos-optimization/VAMOS.git
cd VAMOS
python -m pip install -e ".[dev]"

Use the environment setup above before this command. The default branch may contain work beyond the published release; record the checked-out commit. To inspect the release source, check out the v1.5.0 tag before installing. For ordinary use, prefer the PyPI installation under Core package.