Parallel Genetic Algorithm Library

Description

This is a simple implementation of the genetic algorithm. It is created to find the optimal solution (maximum of a function).

What is implemented?

Selection types:

1. Roulette wheel selection
2. K-Way tournament selection

Mutation:

1. Bit Flip Mutation
2. Swap Mutation

Crossover:

1. One Point Crossover
2. Multi Point Crossover

Usage

Example of usage you can find in the directory test.

#include <iostream>
#include "GeneticAlgorithm.h"

#define PI 3.14159265

using namespace std;

double function(vector<double> args) {
    return exp(-(pow((args[0]-PI),2)+pow((args[1]-PI),2)));
}


int main() {
    long long int elite_number = 10;
    double crossover_percent = 0.6;
    unsigned int number_of_args = 2;
    const function<double(vector<double>)> &func = f;
    double lower_bound = -100;
    double upper_bound = 100;
    double mutation_probability = 0.05;
    unsigned int chromosome_length = 24;
    unsigned int population_size = 100;
    unsigned int threads_number = 4;

    GeneticAlgorithm geneticAlgorithm(
            elite_number, crossover_percent, number_of_args,
            func,
            lower_bound, upper_bound, mutation_probability,
            chromosome_length, population_size,
            threads_number
    );

    unsigned int max_iterations_number = 1000;
    bool running_in_parallel = true;
    
    geneticAlgorithm.iterate_until(max_iterations_number, running_in_parallel);

    geneticAlgorithm.find_max();
}

Contributors:

Kostiantyn Liepieshov, Iryna Kostyshyn, Victoria Yuzkiv