Class: Wallace::Evolver

Inherits:
Object
  • Object
show all
Defined in:
lib/core/evolver.rb

Overview

Steady-state vs. Generational

Instance Method Summary collapse

Constructor Details

#initialize(rng, population, evaluator, termination) ⇒ Evolver

Constructs a new evolver.

Parameters:

  • rng, the random number generator to use for stochastic processes.

  • population, the population to evolve.

  • evaluator, the evaluator to use to determine the fitness of individuals.

  • breeder, the breeder to use to create individuals for successive generations.

  • migrator, the migrator to use to exchange individuals between sub-populations.

  • termination, the termination criteria for the evolution.



13
14
15
16
17
18
19
20
 
# File 'lib/core/evolver.rb', line 13

def initialize(rng, population, evaluator, termination)
  @rng = rng
  @population = population
  @evaluator = evaluator
  @termination = termination
  @evaluator = evaluator
  @state = nil
end

Instance Method Details

#evolve ⇒ Object

Conducts the described evolutionary algorithm until the termination criteria is met. Upon meeting the termination criteria, the best individual found from the entire run is returned.



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
 
# File 'lib/core/evolver.rb', line 25

def evolve

  # Initialise and evaluate the population.
  @population.fresh!(@rng)
  @evaluator.process!(@rng, @population)
  @state = Wallace::State.new(@population)

  # Continue the evolution unless the termination condition has been met.
  until @termination.finished?(@state)
    @migrator.migrate!(@rng, @population) unless @migrator.nil?
    @population.breed!(@rng)
    @evaluator.process!(@rng, @population)
    @state.next!
    #puts "#{@state.best.to_s} = #{@state.best.fitness.value}"
    puts "#{@state.generations}: #{@state.best.fitness.value} (#{@state.best.data.length})"
  end

  # Return the best individual from the run.
  return @state.best

end