Installation¶
PAO currently supports the following versions of Python:
- CPython: 3.7, 3.8, 3.9
Using PIP¶
The standard utility for installing Python packages is pip. You can install the latest release of PAO by executing the following:
pip install pao
You can also use pip to install from the PAO software repository. For example, the master branch can be installed as follows:
python -m pip install https://github.com/or-fusion/pao.git
Using CONDA¶
The conda utility can be used to install the latest release of PAO
using the conda-forge
channel:
conda install -c conda-forge pao
Using GIT¶
PAO can be installed by cloning the PAO software repostory and then directly installing the software. For example, the master branch can be installed as follows:
git clone https://github.com/or-fusion/pao.git
cd pao
python setup.py develop
Conditional Dependencies¶
Both conda and pip can be used to install the third-party packages that are needed to model problems with PAO. We recommend conda because it has better support for optimization solver packages.
PAO intrinsically depends on Pyomo, both for the representation of algebraic problems but also for interfaces to numerical optimizers used by PAO solvers. Pyomo is installed with PAO, but the Pyomo website [PyomoWeb] and GitHub site [PyomoGithub] provide additional resources for installing Pyomo and related software.
PAO and Pyomo have conditional dependencies on a variety of third-party packages, including Python packages like scipy, numpy and optimization solvers. Optimization solvers are particularly important, and a commercial optimizer may be needed to analyze complex, real-world applications.
The following optimizers are used to test the PAO solvers:
- glpk - An open-source mixed-integer linear programming solver
- cbc - An open-source mixed-integer linear programming solver
- ipopt - An open-source interior point optimizer for continuous problems
Additionally, PAO can interface with optimization solvers at NEOS.