Home

Genetic Algorithm for .NET

What is Cormorant?

Cormorant is a framework for C# created to solve optimization problems through genetic algorithms (GAs) in a simple and intuitive way.

For the creation of the classes, functions and objects in general has been used a nomenclature inspired in the field of studio of the biology, avoiding confusing terms or mixtures of abbreviations with the aim of making learning as easy and fast as possible.

Your suggestions and ideas are welcome. Please, send them to support@cormorant.juanexposito.com

Quick Start

Add the CormorantCommon reference to your project.

Add the following using.


    using CormorantCommon;

Create an instance of the Cormorant class.


    var cor = new Cormorant();

Create the genes.


    var genX = new Gene("x", -100d, 100d);
    var genY = new Gene("y", -100d, 100d);

Create the genome and set the genes and the genome.


    var genome = new Genome();
    genome.SetGene(genX);
    genome.SetGene(genY);
    cor.SetGenome(genome);

Set some configurations as the maximum number of generations, the size of the population, the percentage of individuals to be maintained per generation and the percentage of individuals to be mutated per generation.


    cor.SetNumberGenerations(50);
    cor.SetPopulationSize(50);
    cor.SetMaintain(5);
    cor.SetMutation(15);

Optional and just for test purpose: set the solution if it is known to get the absolute and relative error.


    cor.SetSolution(141.42136);

Create the fitness function and set it with the SetFitnessFunction method.


    static double FitnessFunction(Individual individual)
    {
        var genX = individual.Genotype.GetGeneByName("x").Value;
        var genY = individual.Genotype.GetGeneByName("y").Value;

        // Function: f = (x^2 + x^2)^(1/2)
        var result = Math.Sqrt(Math.Pow(genX, 2) + Math.Pow(genY, 2));

        return result;
    }

    cor.SetFitnessFunction(FitnessFunction);

Finally, use the Compute method to compute the algorithm and get the report with the GetReport method.


    cor.Compute();
    Console.WriteLine(cor.GetReport());

The complete code would be the following:


    var cor = new Cormorant();

    // Create genes
    var genX = new Gene("x", -100d, 100d);
    var genY = new Gene("y", -100d, 100d);

    // Create the genome and set the genes and the genome
    var genome = new Genome();
    genome.SetGene(genX);
    genome.SetGene(genY);
    cor.SetGenome(genome);

    // Set some properties
    cor.SetNumberGenerations(50);
    cor.SetPopulationSize(50);
    cor.SetFitnessFunction(FitnessFunction);
    cor.SetMaintain(5);
    cor.SetMutation(15);

    // Optional and just for test purpose: set the solution if it is known to get the absolute and relative error
    cor.SetSolution(141.42136);

    // Compute the algorithm
    cor.Compute();

    // Show the report
    Console.WriteLine(cor.GetReport());

    static double FitnessFunction(Individual individual)
    {
        var genX = individual.Genotype.GetGeneByName("x").Value;
        var genY = individual.Genotype.GetGeneByName("y").Value;

        // Function: f = (x^2 + x^2)^(1/2)
        var result = Math.Sqrt(Math.Pow(genX, 2) + Math.Pow(genY, 2));

        return result;
    }

Output:


    ==================================================
    Parameters
    ==================================================
    Population size: 50
    Generations: 50
    Maintain: 5%
    Mutation: 15%
    Fitness type: Maximize
    Crossover type: OnePoint

    ==================================================
    Best individual by generation
    ==================================================
    Generation 0:   * 136,62290930443666
    Generation 1:   * 136,62290930443666
    Generation 2:   * 136,62290930443666
    Generation 3:   * 136,62290930443666
    Generation 4:   ********************* 140,29694450867675
    Generation 5:   ********************* 140,29694450867675
    Generation 6:   ********************* 140,29694450867675
    Generation 7:   ********************* 140,29694450867675
    Generation 8:   ********************* 140,29694450867675
    Generation 9:   ********************* 140,29694450867675
    Generation 10:  ********************* 140,29694450867675
    Generation 11:  ********************* 140,29694450867675
    Generation 12:  ********************* 140,29694450867675
    Generation 13:  ********************* 140,29694450867675
    Generation 14:  ********************* 140,29694450867675
    Generation 15:  ********************* 140,29694450867675
    Generation 16:  ********************* 140,29694450867675
    Generation 17:  ********************* 140,29694450867675
    Generation 18:  ********************* 140,29694450867675
    Generation 19:  ********************* 140,29694450867675
    Generation 20:  ********************* 140,29694450867675
    Generation 21:  ********************* 140,29694450867675
    Generation 22:  ********************* 140,29694450867675
    Generation 23:  ************************** 141,20236485076495
    Generation 24:  ************************** 141,20236485076495
    Generation 25:  ************************** 141,20236485076495
    Generation 26:  ************************** 141,20236485076495
    Generation 27:  ************************** 141,20236485076495
    Generation 28:  ************************** 141,20236485076495
    Generation 29:  ************************** 141,20236485076495
    Generation 30:  ************************** 141,20236485076495
    Generation 31:  ************************** 141,20236485076495
    Generation 32:  ************************** 141,20236485076495
    Generation 33:  ************************** 141,20236485076495
    Generation 34:  ************************** 141,20236485076495
    Generation 35:  ************************** 141,20236485076495
    Generation 36:  ************************** 141,20236485076495
    Generation 37:  ************************** 141,20236485076495
    Generation 38:  ************************** 141,20236485076495
    Generation 39:  ************************** 141,20236485076495
    Generation 40:  ************************** 141,20236485076495
    Generation 41:  ************************** 141,20236485076495
    Generation 42:  ************************** 141,20236485076495
    Generation 43:  ************************** 141,20236485076495
    Generation 44:  ************************** 141,20236485076495
    Generation 45:  ************************** 141,22465435877032
    Generation 46:  *************************** 141,25388443440247
    Generation 47:  *************************** 141,25388443440247
    Generation 48:  *************************** 141,25388443440247
    Generation 49:  *************************** 141,25388443440247
    Generation 50:  *************************** 141,25388443440247

    ==================================================
    Best individual
    ==================================================
    Fitness: 141,25388443440247
    Genotype:
    - Gene x, Value: -99,85396962347866
    - Gene y, Value: -99,90918185152418

    ==================================================
    Errors
    ==================================================
    Absolute error: 0,16747556559752752
    Relative error: 0,11842310496627066 %