v1.4.1

• Fixes assignment error when setting Direction
• Adds script to invoke CLI using platypus instead of python -m platypus

v1.4.0

• Adds complete API docs
• Adds Direction enum to specify optimization direction. E.g., replace Problem.MAXIMIZE with Direction.MAXIMIZE.
• Adds save_state and load_state methods, allowing one to resume a run from the state file.
• Adds CLI (see python -m platypus —help)
• Adds extensions (a way to add or change the behavior of an algorithm using extension points)

v1.3.1

Restores the default_variator, default_mutator, and nondominated_cmp methods that were removed in 1.3.0, potentially breaking any code directly calling these methods. These methods are deprecated and will be removed in a future release, so we strongly recommend updating affected code.

1. Replace default_variator(...) with PlatypusConfig.default_variator(...)
2. Replace default_mutator(...) with PlatypusConfig.default_mutator(...)
3. Replace nondominated_cmp with nondominated_sort_cmp

v1.3.0

Updates

• Minimum supported Python version increased to 3.8
• Adds I/O methods for saving and loading populations
• Clean up required dependencies (no longer require numpy, use pip install platypus-opt[full] to get all optional dependencies)

Breaking Changes

This version moved / renamed the default_variator, default_muator, and nondominated_cmp methods. Since this will break any code depending on these methods, they are being restored in version 1.3.1 but are being marked as deprecated. Please update your code as these will be removed in a future version.

v1.2.0

Minor updates since previous release.

v1.1.0

• Upgrade to Python 3
• Remove six dependency, used to bridge compatibility between Python 2 and 3
• Improvements to CI and deployment pipelines

Version 1.0.4

Bump version to 1.0.4

Version 1.0.2

Fixes a bug in how weight_generator was used in MOEA/D.

Version 1.0.1

Adds MANIFEST file to include all docs and examples.

Version 1.0.0

Change name for PyPI