maingopy 0.5.0

A Python package for using MAiNGO - McCormick-based Algorithm for mixed-integer Nonlinear Global Optimization

maingopy - Python interface for MAiNGO

Maingopy is the Python interface for MAiNGO, the McCormick-based Algorithm for mixed-integer Nonlinear Global Optimization. MAiNGO is a deterministic global optimization solver for nonconvex mixed-integer nonlinear programming problems. For more information on MAiNGO, please visit the MAiNGO website. The open source version of MAiNGO is available on our GitLab page. The documentation of MAiNGO is available here.

Obtaining maingopy

Maingopy can either be obtained as a source of binary distribution via PyPI or built from source via the git repository.

To obtain it via PyPI, run

$ pip install maingopy

This will typically get you the binary distribution of the maingopy package that contains a pre-compiled version of MAiNGO along with its Python bindings, as well as an extension module for MeLOn, which contains machine learning models for use in optimization problems to be solved by MAiNGO. Note that the pre-compiled version of MAiNGO contained in this package does not allow the use of the optional closed-source subsolvers CPLEX or KNITRO, even if they are installed on your system.

To build maingopy (and melonpy) from source along with MAiNGO, obtain the code from our GitLab page and follow the instructions provided there. This also allows using CPLEX or KNITRO if you have them installed.

Using maingopy

Maingopy provides Python bindings (enabled by pybind11) for the C++ API of MAiNGO. Details on how to use it are available in the documentation of MAiNGO. Example problems can be found in the examples directory in the MAiNGO repository.