This is a pre-production deployment of Warehouse, however changes made here WILL affect the production instance of PyPI.
Latest Version Dependencies status unknown Test status unknown Test coverage unknown
Project Description
# Python-TinyEvolver
### A simple, tiny engine for creating genetic algorithms

TinyEvolver is a framework for creating genetic algorithms written in pure python.
It aims to let you write sensible evolutionary algorithms in as few steps as possible
using a prototype system to extrapolate generation, mutation and mating of individuals
from a simple example.

TinyEvolver was developed for scientists and researchers who want to utilize genetic
algorithms in models and applications, but not necessarily become researchers in
genetic/evolutionary algorithms themselves. You define the things that are really
unique to your problem and TinyEvolver does the rest.

The source code for TinyEvolver is inspired, at least in part, by the [DEAP](
module, but we've made conscious decisions to tailor our module towards simplicity and lightness.
But simple doesn't mean featureless - individuals can have genes of mixed type, populations
can be generated on the fly or from old data, and one can evolve many populations at
once with multiprocessing.

## Installation
Installation requires Python 2.6+ or Python 3.4+.

To install from source, simply clone into a directory, then from that directory run
python install
or, if you'd prefer to be able to edit the installed code yourself:
python develop

## Example
from tinyevolver import Population

prototype = [False for _ in range(100)]
p = Population(prototype=prototype, gene_bounds=None, fitness_func=sum)



## Tips
The majority of the work in constructing an evolutionary algorithm in TinyEvolver is the fitness function - and this is where the majority of the work is done by the CPU. You can thus speed up your code by speeding up the fitness function, whether that be by outsourcing to NumPy, writing C extensions, or simply making your function more efficient. Since TinyEvolver is written in pure Python, you could also run it under [PyPy](

The best way to discover TinyEvolver's features is through the iPython interactive interpreter - you can enter `Foo.` followed by the `tab` key to see possible completions of Foo, and `Foo?` to view its signature and docstrings.

## Documentation
TinyEvolver contains 3 classes: Individual, Population and IslandModel. A Population is a collection of Individuals and an IslandModel is a collection of Populations - both of these classes have methods for evolving with all variables having sensible defaults.

### Individual
Users should not need to create an instance of this class directly.

- `individual.genes` a 1D-array or flat list of genes.
- `` the individual's fitness - may or may not be present.
- `individual.valid` is True only if `` is present.

- Individuals have many of the methods of lists: you can get/set their genes with indices or slices, iterate over them, put them into `len`, copy them, and put them into any other Python function requiring only these.

### Population
Create an instance with `Population(prototype, gene_bounds, fitness_func)`, where
- `prototype` is a flat list of booleans, integers and floats whose types individuals' genes should have (namely boolean, float or integer).
- `gene_bounds` is either None or a list of lower/upper bounds for the genes.
- `fitness_func` takes a flat list of genes and returns a numeric value representing the individual's fitness.

- `` the individual with the highest fitness.
- `population.individuals` the full list of individuals in the population.

- Populations have many of the methods of lists: you can get/set their individuals with indices or slices, iterate over them, put them into `len`, copy them, or put them into any other Python function requiring only these.
- `population.populate([popsize, base_population])` if no `base_population` is passed then this will generate the required number of individuals for the population using its `prototype` and `gene_bounds`. If a family of list-like objects is passed as a `base_population` then the population is populated with these instead.
- `population.evolve([ngen, matepb, mutpb, indpb, scoping, tournsize, verbose])` this should only be called after the class has been populated. It evolves `ngen` generations, where individuals have a probability `matepb` of mating, `mutpb` of mutating. `indpb` controlls the variability of an individual's genes upon mutation. Fitest individuals are selected from random tournaments of size `tournsize`. If `scoping` is positive then the amount by which floats are able to mutate decreases from one generation to the next - honing in upon parameters. Set `verbose` to False to avoid printing details of the evolution.

### IslandModel
Create an IslandModel instance with `IslandModel(poplist)` where `poplist` is a list of `Population` objects.

- `` the best individual from all the individual populations
- `islandmodel.islands` a list containg the class' populations

- `islandmodel.amalg_pop()` this returns the islands amalgamated into a single large population
- `islandmodel.select_pop()` this selects a population from across the islands whose size is that of a single island
- `islandmodel.evolve([ngen, matepb, mutpb, indpb, scoping, tournsize, verbose, mig_freq])` this evolves all the islands, with individuals migrating between islands every `mig_freq` generations. See the `evolve` method for the `Population` class.
- `islandmodel.multi_evolve([ngen, matepb, mutpb, indpb, scoping, tournsize, verbose, mig_freq])` this is the same as the `evolve` method, but uses multiprocessing.
Release History

Release History


This version

History Node

TODO: Figure out how to actually get changelog content.

Changelog content for this version goes here.

Donec et mollis dolor. Praesent et diam eget libero egestas mattis sit amet vitae augue. Nam tincidunt congue enim, ut porta lorem lacinia consectetur. Donec ut libero sed arcu vehicula ultricies a non tortor. Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Show More

Download Files

Download Files

TODO: Brief introduction on what you do with files - including link to relevant help section.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
tinyevolver-0.1-py2.py3-none-any.whl (13.1 kB) Copy SHA256 Checksum SHA256 py2.py3 Wheel Oct 15, 2015
tinyevolver-0.1.tar.gz (8.6 kB) Copy SHA256 Checksum SHA256 Source Oct 15, 2015

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS HPE HPE Development Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting