Configuration

Genex offers a number of configuration options. This is a comprehensive list.

To adjust radiation, crossover, and mutation rates, please see the Customization guide.

Configuration options are provided after the use Genex line in your module.

defmodule MyGeneticAlgorithm do
    use Genex, crossover_type: :two_point, mutation_type: :invert
    ...
end

Defaults

• :population_size: 100
• :crossover_type: :single_point
• :parent_selection: :natural
• :survivor_selection: :natural
• :mutation_type: :scramble
• :minimize: false

crossover_type

This option determines what kind of Crossover to use. For a detailed explanation of each Crossover method, see the Crossover guide.

• :single_point
• :two_point
• :uniform (requires :uniform_crossover_rate parameter set)
• :blend (requires :alpha parameter set)
• :simulated_binary (requires :eta parameter set)
• :messy_single_point
• :davis_order

parent_selection and survivor_selection

This option determines what kind of Selection to use. For a detailed explanation of each Selection method, see the Selection guide.

• :natural
• :random
• :worst
• :roulette
• :tournament (requires :tournsize parameter set)
• :stochastic

mutation_type

This option determines what kind of mutation to use. For a detailed explanation of each Mutation method, see the Mutation guide.

• :bit_flip
• :invert
• :scramble
• :gaussian
• :uniform_integer (requires upper_bound and lower_bound parameters set)
• :polynomial_bounded (requires upper_bound, lower_bound and eta parameters set)
• :none

minimize

This option tells Genex if you want to minimize or maximize fitness. When set to true, Genex will sort population fitness in ascending order. The "strongest" chromosome will be the one with the lowest fitness.