A Python platform to perform parallel computations of optimisation tasks (global and local) via the asynchronous generalized island model.
This new pygmo release includes fixes for Python 3.12 and adds batch fitness evaluation capabilities to the CMAES algorithm.
This is a minor release featuring several documentation fixes, and adding pre-built wheels for 64-bit Linux ARM (aarch64
), courtesy of @ccrutchf.
Full changelog: https://esa.github.io/pygmo2/changelog.html
Due to the large number of requests we have refreshed and restored the binary wheel creation for the linux platform.
A new meta-problem allows to fix some of the decision vector components and reduce dimensionality.
Added a generational version of MOEA-D with batch fitness evaluation support.
Full changelog: https://esa.github.io/pygmo2/changelog.html
pygmo 2.18.0 adds official support for non-x86 processors, specifically 64-bit ARM and PowerPC. Automated builds on these architectures have been added to the CI pipeline, courtesy of travis and circleci.
This release includes various changes in the serialisation of the main classes. As a result, it's not possible in pygmo 2.18.0 to unpickle from archives produced with earlier versions of pygmo.
pygmo now requires CMake >= 3.17 when building from source. A compilation failure in debug mode with Python >= 3.9 has been fixed.
This is a minor release containing a few build system fixes, and bumping up the minimum required pybind11 version to 2.6.
As usual, the full changelog is available here:
The major new feature in this new pygmo release is an algorithm wrapper around the SciPy optimizers, developed by @mlooz
Additionally, like pagmo 2.16.0, pygmo now requires a C++17-capable compiler when installing from source.
The full changelog, as usual, is available here: