Crate chromosome

Chromosome

Utilities for genetic algorithms

Example

use core::{
    iter::Sum,
    ops::{Mul, Sub},
};

use chromosome::Fitness;
use chromosome::{Chromosome, DefaultSimulator, FitnessSelector, Simulator};
use rand_pcg::Pcg64;
use rand_seeder::Seeder;

struct DiophantusEquation<'a, 'b, T> {
    coefficients: &'a Vec<T>,
    result: &'b T,
}

impl<'a, 'b, T> DiophantusEquation<'a, 'b, T> {
    fn new(coefficients: &'a Vec<T>, result: &'b T) -> Self {
        DiophantusEquation {
            coefficients: coefficients,
            result: result,
        }
    }
}

impl<'a, 'b, T: Mul<Output = T> + Sum + Sub<Output = T> + Into<f64> + Clone> Fitness
    for DiophantusEquation<'a, 'b, T>
{
    type Value = T;
    fn fitness(self: &Self, chromosome: &Chromosome<T>) -> T {
        (0..usize::min(self.coefficients.len(), chromosome.genes.len()))
            .map(|i| self.coefficients[i].clone() * chromosome.genes[i].clone())
            .sum::<T>()
            - self.result.clone()
    }

    fn is_ideal_fitness(&self, fitness: Self::Value) -> bool {
        fitness.into() == 0_f64
    }
}

#[test]
fn diophantus_equation() {
    let mut rng: Pcg64 = Seeder::from([23, 87, 85]).into_rng();
    let coefs = vec![2_i32, 23, 54, 1];
    let equation = DiophantusEquation::new(&coefs, &2);

    let sim_result = DefaultSimulator::new(vec![1; 4], 0.09, 10000).simulate(
        vec![
            Chromosome::new_random(equation.coefficients.len(), 0_i32..10, &mut rng),
            Chromosome::new_random(equation.coefficients.len(), 0_i32..10, &mut rng),
        ],
        FitnessSelector::from(equation),
        &mut rng,
    );

    assert!(sim_result.is_some())
}

Structs

Chromosome

Chromosome contains genes and provide genetic operations on them

DefaultSimulator

DefaultSimulator make default simulation

FitnessSelector

FitnessSelector selects chromosomes by fitness values

SimulationIter

SimulationIter provides Simulating genetics by iterating. each next() call produces next generation. .collect() gives all generations up to final ideal one

Traits

Fitness

Fitness trait provides interface for finging fitness value for chromosome

Selector

Genetic selector

Simulator

Simulator makes blocking simulation of genetics returns Chromosome if simulation successful and solution found and None if no solution

Superposition

Functions

map_random_pairs