Implementation:Haifengl Smile Crossover

Knowledge Sources	Haifengl_Smile
Domains	Genetic Algorithms, Optimization, Evolutionary Computation
Last Updated	2026-02-08 22:00 GMT

Overview

Crossover is an enum defining the types of crossover operations available for bit-string chromosomes in the Smile genetic algorithm package.

Description

The Crossover enum encapsulates three classical crossover strategies used in genetic algorithms operating on binary string representations. Each enum constant implements the abstract apply method to produce two offspring from two parent BitString chromosomes.

The three supported crossover strategies are:

• SINGLE_POINT -- selects one random crossover point. The binary string from the beginning to the crossover point is copied from the first parent, and the rest from the second parent. The second offspring receives the complementary segments.

• TWO_POINT -- selects two random crossover points. The segment between the two points is swapped between parents. This is the default crossover strategy used by BitString.

• UNIFORM -- each bit is independently and randomly copied from either parent with equal probability (50/50). This provides maximum mixing of genetic material.

All strategies produce exactly two offspring (son and daughter) from two parents (father and mother), maintaining population size during evolution.

Usage

Use the Crossover enum when constructing BitString chromosomes to specify the desired crossover strategy. The default is TWO_POINT.

Code Reference

Source Location

• Repository: Haifengl_Smile
• File: base/src/main/java/smile/gap/Crossover.java
• Lines: 1-136

Signature

public enum Crossover {

    SINGLE_POINT {
        @Override
        public BitString[] apply(BitString father, BitString mother) { ... }
    },

    TWO_POINT {
        @Override
        public BitString[] apply(BitString father, BitString mother) { ... }
    },

    UNIFORM {
        @Override
        public BitString[] apply(BitString father, BitString mother) { ... }
    };

    /**
     * Returns a pair of offsprings by crossovering parent chromosomes.
     */
    public abstract BitString[] apply(BitString father, BitString mother);
}

Import

import smile.gap.Crossover;

I/O Contract

Inputs

Name	Type	Required	Description
father	BitString	Yes	The first parent chromosome.
mother	BitString	Yes	The second parent chromosome. Must have the same bit string length as the father.

Outputs

Name	Type	Description
result	BitString[]	An array of two offspring BitString chromosomes created by combining genetic material from the parents.

Usage Examples

Basic Usage

import smile.gap.BitString;
import smile.gap.Crossover;
import smile.gap.Fitness;

// Define a simple fitness function
Fitness<BitString> fitness = chromosome -> {
    byte[] bits = chromosome.bits();
    int count = 0;
    for (byte b : bits) count += b;
    return count;
};

// Create chromosomes with single-point crossover
BitString parent1 = new BitString(32, fitness, Crossover.SINGLE_POINT, 0.9, 0.01);
BitString parent2 = new BitString(32, fitness, Crossover.SINGLE_POINT, 0.9, 0.01);

// Directly apply crossover
BitString[] offspring = Crossover.SINGLE_POINT.apply(parent1, parent2);
System.out.println("Offspring 1: " + offspring[0]);
System.out.println("Offspring 2: " + offspring[1]);

Comparing Crossover Strategies

import smile.gap.BitString;
import smile.gap.Crossover;
import smile.gap.Fitness;

Fitness<BitString> fitness = ch -> {
    byte[] bits = ch.bits();
    int score = 0;
    for (byte b : bits) score += b;
    return score;
};

// Create populations with different crossover strategies
for (Crossover strategy : Crossover.values()) {
    BitString[] pop = new BitString[30];
    for (int i = 0; i < pop.length; i++) {
        pop[i] = new BitString(64, fitness, strategy, 0.9, 0.01);
    }
    // Use with GeneticAlgorithm...
}

Related Pages

• Implementation:Haifengl_Smile_BitString
• Implementation:Haifengl_Smile_Chromosome
• Implementation:Haifengl_Smile_GeneticAlgorithm
• Implementation:Haifengl_Smile_GA_Parent_Selection