contrebis/blue-jeans

There is no license information available for the latest version (dev-master) of this package.

A framework for implementing genetic algorithms

dev-master 2013-12-08 02:28 UTC

Requires

• php: >=5.4.0
• doctrine/collections: *

Requires (Dev)

• phpunit/phpunit: *

Unknown License

• Robert Churchill

README

Introduction

This component is/will be a framework for solving optimisation problems using genetic algorithms.

The idea is you create a random pool of 'genomes'. Your genome class extends Genome and must interpret your bit string into a solution to your problem and calculate a fitness value for the solution. Good solutions are more likely to survive into the next generation.

Configurable options are:

• Elitism: how many of the fittest genomes survive to the next generation intact.
• Mutation rate
• Crossover rate
• Population size

An indicator of diversity can be calculated to show how much entropy remains in the system. If wanted, the mutation rate could vary to keep the diversity above a certain threshold.

TODO

• Better test coverage (and remove whatever bugs are still lurking in there)
• Some documentation
• Optimise list collection class
• Other performance optimisations
• Add a license