METAOPTIM

metaoptim is a Python package for metaheuristic optimization algorithms. It also features various benchmark functions that can be used to test the performance of the algorithms.

Installation

Ensure that you have Python 3.9 or higher installed. Then, install the package using pip:

pip3 install open-metaoptim

Usage

Configuration

This package has the ability to accelerate the metaheuristic algorithms using numba. To control the use of numba you can control the following flags in the metaoptim.config module:

import multiprocessing
_numba_cache = True                                 # Set to True/False to enable/disable Numba cache
_numba_parallel = False                             # Set to True/False to enable/disable Numba parallel
_numba_nopython = True                              # Set to True/False to enable/disable Numba nopython
_numba_nogil = True                                 # Set to True/False to enable/disable Numba nogil
_numba_num_threads = multiprocessing.cpu_count()    # Set to the number of threads to use for Numba
_disable_jit = True                                 # Set to True/False to enable/disable Numba JIT

The aforementioned flags are set to their default values. You can change them to your liking by importing the config module and setting the flags to the desired values. e.g.:

import metaoptim.config as config
config._numba_cache = False

However, please ensure that you import any metaheuristic algorithms after you have set the flags. Otherwise, the flags will not take effect.

Finally, GPU acceleration is currently not supported, but will be added in the future.

Recommendations

Numba is a great tool for accelerating Python code. However, it is not always the best choice. For example, if you are using a metaheuristic algorithm that is already fast, then you may not want to use Numba. In such cases, you can disable Numba by setting the _disable_jit flag to True. You should use Numba acceleration when the metaheuristic algorithm is slow, working in high dimensions, or when you are using a large population size.

Algorithms

The package currently features the following algorithms:

• Global Best Particle Swarm Optimization

Please refer to the class docstrings for more information.

Benchmark Functions

• Ackley (Dimension: N)
• Beale (Dimension: 2)
• Booth (Dimension: 2)
• Bukin N. 6 (Dimension: 2)
• Cross-in-Tray (Dimension: 2)
• Dixon-Price (Dimension: N)
• Drop-Wave (Dimension: 2)
• Easom (Dimension: 2)
• Eggholder (Dimension: 2)
• Goldstein-Price (Dimension: 2)
• Griewank (Dimension: N)
• Himmelblau (Dimension: 2)
• Holder Table (Dimension: 2)
• Levy (Dimension: N)
• Levy N. 13 (Dimension: 2)
• Matyas (Dimension: 2)
• McCormick (Dimension: 2)
• Rastrigin (Dimension: N)
• Rosenbrock (Dimension: N)
• Schaffer N. 2 (Dimension: 2)
• Schaffer N. 4 (Dimension: 2)
• Sphere (Dimension: N)
• Styblinski-Tang (Dimension: N)
• Three-Hump Camel (Dimension: 2)

Package Structure

The package is structured as follows:

• metaoptim: The package directory.
  • config.py: The configuration file.
  • bench_func.py: Contains the benchmark functions.
  • pso: The Particle Swarm Optimization algorithms package.
    • pso.py: The PSO parent class.
    • gbest_pso.py: The Global Best Particle Swarm Optimization algorithm (GBestPSO). Features an inertia weight, a cognitive and a social component.

Contributing

Guidelines for Contributing

NEW METAHEURISTICS MUST INCLUDE NUMBA ACCELERATION

1. In the description of your pull request, please explain clearly what does it implement or fix and your changes. Where possible, you should give examples in the description of the pull request. In cases that the pull request is about a code speedup, report a reproducible example and quantify the speedup.

2. Give your pull request a helpful title that summarises what your contribution does.

3. Write unit tests for your code in the test directory and make sure the existing tests are up-to-date. Use pytest to run the tests:

pytest

4. Make sure that there a no stylistic issues in files you have edited using pycodestyle:

pycodestyle <FILE_PATH>

where <FILE_PATH> is the path to the file you want to check.

To check all files in the package, run:

pycodestyle src/*

5. Make sure your code is properly commented and documented using reStructuredText docstrings and additional comments where necessary. Each public method needs to be documented as the existing ones.

Development Dependencies

To install the development dependencies, run the following command:

pip3 install metaoptim[dev]