Skip to main content

Cartesian genetic programming (CGP) in pure Python.

Project description

https://badge.fury.io/py/hal-cgp.svg https://img.shields.io/badge/python-3.6-red.svg https://img.shields.io/badge/python-3.7-red.svg https://img.shields.io/badge/python-3.8-red.svg https://img.shields.io/badge/License-GPLv3-blue.svg https://github.com/Happy-Algorithms-League/hal-cgp/actions/workflows/tests.yaml/badge.svg http://www.mypy-lang.org/static/mypy_badge.svg https://img.shields.io/badge/code%20style-black-000000.svg https://coveralls.io/repos/github/Happy-Algorithms-League/python-gp/badge.svg?branch=master https://readthedocs.org/projects/hal-cgp/badge/?version=latest

Cartesian genetic programming (CGP) in pure Python.

hal-cgp is an extensible pure Python library implementing Cartesian genetic programming to represent, mutate and evaluate populations of individuals encoding symbolic expressions targeting applications with computationally expensive fitness evaluations. It supports the translation from a CGP genotype, a two-dimensional Cartesian graph, into the corresponding phenotype, a computational graph implementing a particular mathematical expression. These computational graphs can be exported as pure Python functions, NumPy-compatible functions (Walt et al., 2011), SymPy expressions (Meurer et al., 2017) or PyTorch modules (Paszke et al., 2019).

The library implements a mu + lambda evolution strategy (Beyer and Schwefel, 2002) to evolve a population of individuals to optimize an objective function.

Design decisions/use cases

We designed hal-cgp for optimization problems in which individual fitness evaluations are computationally expensive. The library is hence not optimized for high performance, but rather puts ease of use and extensibility first. Furthermore we take steps to reduce the number of redundant fitness evaluations, for example by avoiding reevaluating parents at the beginning of each episode and providing a convenient decorator to cache results on disk. If for your use case individual fitness evaluations are fast and the performance of the library itself becomes a relevant factor, you may want to check out https://github.com/darioizzo/dcgp or http://www.cgplibrary.co.uk/files2/About-txt.html.

CGP Sketch

Figure from Jordan, Schmidt, Senn & Petrovici, “Evolving to learn: discovering interpretable plasticity rules for spiking networks”, arxiv:2005.14149.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

hal-cgp-0.3.0.tar.gz (61.8 kB view hashes)

Uploaded source

Built Distribution

hal_cgp-0.3.0-py3-none-any.whl (46.9 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page