A fast differential evolution module
PyFDE is an implementation of differential evolution for Python 3. Its main focus is ease of use and performance. The core optimization procedure was implemented in Cython for performance.
Currently, it implements the classic DE/rand/1/bin scheme and the JADE variant. pyfde also provides a fast random number generator.
The documentation is available at http://pythonhosted.org/PyFDE .
To install PyFDE from the source package, run:
python setup.py install
PyFDE targets Python 3.4 and depends on NumPy at compile/run time. Cython is only needed when changing the library itself (needed for generating the C modules). Optionally, one can implement the fitness function in Cython for performance too.
The project is licensed under the MIT license.