Struct jeans::Settings

pub struct Settings {
    pub population_size: u32,
    pub number_of_generations: u32,
    pub crossover_probability: f64,
    pub mutation_probability: f64,
    pub verbose: bool,
    pub elitism: bool,
    pub elitism_fraction: f64,
    pub maximize_fitness: bool,
    pub number_of_dimensions: u32,
    pub upper_bound: Vec<f64>,
    pub lower_bound: Vec<f64>,
    pub fitness_function: Box<dyn FnMut(Vec<f64>) -> f64>,
}

A settings object for storing all of the settings we might care about for a GA.

You should usually instantiate this using the default method. All member variables of this struct are public, which makes editing this easy. In most cases, simply reassign the member variables:

let mut set = jeans::Settings::default();
set.elitism = false;

The most notable exception is for the fitness_function field. To set that, you have to use a special method:

let mut set = jeans::Settings::default();
set.set_fitness_function(| x: Vec<f64> | x[0] + x[1])

Fields

population_size: u32

The size of the population

number_of_generations: u32

The number of generations

crossover_probability: f64

The crossover probability

mutation_probability: f64

The probability of mutation

verbose: bool

A bool indicating whether or not code should run in verbose mode

elitism: bool

A bool indicating whether or not to implement elitism

elitism_fraction: f64

The fraction of each generation that should be carried forward via elitism

maximize_fitness: bool

Whether the package should maximize fitness (true) or minimize fitness (false)

number_of_dimensions: u32

The number of dimensions of the problem space

upper_bound: Vec<f64>

The upper bounds for each dimension of the problem space

lower_bound: Vec<f64>

The lower bounds for each dimension of the problem space

fitness_function: Box<dyn FnMut(Vec<f64>) -> f64>

The fitness function used to evaluate how good solutions are

Implementations

impl Settings

pub fn set_fitness_function<F: 'static + FnMut(Vec<f64>) -> f64>(
    &mut self,
    f: F
)

This functions allows you to set the fitness function

For instance, you can use one of hte built-in function found here:

let mut set = jeans::Settings::default();
set.set_fitness_function(jeans::functions::sphere);

Or you can use a lambda to implement your own:

let mut set = jeans::Settings::default();
set.set_fitness_function(| x | x[0] + x[1]);

Or you can define and use a completely new function

fn fitness_whole_pizza(x: Vec<f64>) -> f64 {
    let mut fx = 0f64;
    for i in 0..x.len() {
        fx *= x[i]
    }
    fx
}
let mut set = jeans::Settings::default();
set.set_fitness_function(fitness_whole_pizza);

Trait Implementations

impl Default for Settings

Default values for all settings

fn default() -> Self

Returns the "default value" for a type.