Skip to content
/ talgene Public

Simple abstraction layer over genetic algorithms implementations

License

Notifications You must be signed in to change notification settings

nin93/talgene

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

44 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Talgene

A simple programming abstraction layer over genetic algorithms implementations.

Installation

  1. Add the dependency to your shard.yml:

    dependencies:
      talgene:
        github: nin93/talgene
  2. Run shards install

Usage

require "talgene"

Genome

Start by defining your genetic representation for your model. In the following example we implement a solution for the knapsack problem.

An abstract class Talgene::Genome is provided for this purpose, requiring you to implement all the methods needed. Within each model a fitness function is expected to be defined to evaluate the solution domain:

record Item, value : Float64, weight : Float64 do
  property? inside : Bool = false
end

class Knapsack < Talgene::Genome(Item)
  def initialize(@genes : Array(Item), @max_weight : Float64)
  end

  def fitness
    weight = 0.0

    reduce 0.0 do |fitness, item|
      unless item.inside?
        fitness
      else
        if (weight += item.weight) > @max_weight
          break 0.0
        else
          fitness + item.value
        end
      end
    end
  end
end

One could implement custom, yet classic, rules for genetic recombination such as crossover and mutation functions. For example, using our Knapsack model from above:

class Knapsack < Talgene::Genome(Item)
  # Storing a mutation rate as well
  def initialize(@genes : Array(Item), @max_weight : Float64, @mutation_rate : Float64)
  end

  def cross(other : Knapsack) : Knapsack
    new_genes = Talgene::Crossable.single_point_cross(@genes, other.genes).sample

    Knapsack.new new_genes, @max_weight, @mutation_rate
  end

  def mutate : Knapsack
    new_genes = map do |item|
      new_item = item.dup

      if @mutation_rate > rand
        new_item.inside = !new_item.inside?
      end

      new_item
    end

    Knapsack.new new_genes, @max_weight, @mutation_rate
  end
end

Generation

We now need to define the rules to perform a selection among competing individuals within a population and start a new generation. This is done by inheriting from the Talgene::Generation abstract class and by implementing an advance method:

class Generation < Talgene::Generation(Knapsack)
  def advance : Generation
    # Avoid recombination with self
    other_bucket = population.reject do |knapsack|
      knapsack.same? fittest
    end

    new_population = Array.new population.size do
      fittest.cross(other_bucket.sample).mutate
    end

    Generation.new new_population
  end
end

System

Talgene::System takes care to iterate through generations. We load the generation zero:

# Initialize your generation zero
population_zero = [...] of Knapsack
generation_zero = Generation.new population_zero

sys = Talgene::System.new generation_zero, max_advances: 100

# Use `include_first: true` to include the generation zero
sys = Talgene::System.new generation_zero, max_advances: 100, include_first: true

Since Talgene::System includes the Iterator module, a set of convenient methods are provided such as max_by, skip_while, select, each_cons.

# The fittest among all generations is easy to find with `max_of`.
fittest_ever = sys.max_of &.fittest
fittest_ever.fitness # => 26.5

Optionally, a Talgene::System can be initialized with a rule declaring when an evolution process should be ended in advance, useful in those cases in which we would expect a good enough individual beforehand.

For instance:

sys = Talgene::System.new generation_zero, max_advances: 100 do
  stop_on do |current|
    current.best_fitness > 26
  end
end

# Consume the iterator
sys.size # => 4
sys.next # => Iterator::Stop::INSTANCE

A full example can be found in the examples folder.

Contributing

  1. Fork it (https://github.com/nin93/talgene/fork)
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Add some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create a new Pull Request

Contributors

About

Simple abstraction layer over genetic algorithms implementations

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published