fengrk / sapso

Phyton implementation of the SAPSO algorithm

Python SAPSO and PSAPSO

Python implemetations (sequential and parallel) of the semi autonomous particle swarm optmizer (sapso)

Prerequisites

numpy

Optmizing a function:

1- Define your parameters of optmization in a dictionary as specified in the example

2- Call the optmizer of your choice passing those parameters:

   from psapso.optmizer import sapso
   sapso(parameters)

Description of files:

• /sapso:

  SAPSO basic implementation, allows to chose between sequential and parallel gradient calculation

• /psapso:

  Parallel SAPSO beta implementation

• /psapso slow_info_exchange:

  Parallel SAPSO based on slow informaitone exchange beta implementation

• /scripts:

  Scripts used to benchmark the optmizers

• test_functions.py:

  A module composed of mathematical functions for optmization tests

• example_parameters.py:

  An example on how to build a parameters dictionary that must be passed to the optmizer

• /scripts/example_file_name.py.lprof:

  Binary file contaning a line_profiler (detailed information on execution time per line) of the optmizer main function (for execution time improvement purposes)

Authors

• Marcelo Freitas - marcelofa

Acknowledgments

• Reginaldo Santos - initial sapso paper and implementation
• Abner Cardoso - initial studies on parallel sapso and implementation