Genetic Algorithms in C++

Overview

gapp is a C++ library for solving single- and multi-objective optimization problems using genetic algorithms. The main features of the library include:

• Simple usage with predefined genetic operators, encoding types, and algorithms
• Includes implementations of several commonly used encoding types (binary, real, permutation, integer), with support for mixed and custom encoding types
• Includes implementations of commonly used crossover and mutation methods for each builtin encoding type, with support for custom genetic operators
• Support for both single- and multi-objective problems, constrained optimization problems, and support for both fixed and variable length solution encodings
• Builtin support for metrics to observe the evolution process, including implementations of common metrics like mean, minimum, and maximum fitness values in each generation
• Reproducible results across platforms, including with multithreaded execution
• Efficient, multithreaded execution with minimal overhead for the GAs, while also minimizing the number of objective function evaluations

Usage example

#include <gapp/gapp.hpp>
#include <iostream>

using namespace gapp;

class SinX : public FitnessFunction<RealGene, 1> 
{
    FitnessVector invoke(const Candidate<RealGene>& x) const override { return { std::sin(x[0]) }; }
};

int main()
{
    auto solutions = RCGA{}.solve(SinX{}, Bounds{ 0.0, 3.14 });

    std::cout << "The maximum of sin(x) in [0.0, 3.14] is at x = " << solutions[0].chromosome[0];
}

Possible console output:

The maximum of sin(x) in [0.0, 3.14] is at x = 1.57079

Requirements

The following are needed for building and using the library:

• C++20 compiler (gcc 11, clang 15, msvc 14.30 or later)
• CMake 3.21 or later
• Catch2 3.3 or later (optional, only needed for the tests)

Installing the library

See install-guide.md for a more detailed installation and usage guide.

The library uses CMake as its build system. Assuming you have all of the requirements listed above, the steps for installing the library in Release config are:

git clone https://github.com/KRM7/gapp.git --branch v1.0.0
cd gapp/build
cmake .. -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=OFF
cmake --build . --config Release
sudo cmake --install . --config Release

Alternatively, you can also use the install script that is provided with the library, which will install all available configurations:

git clone https://github.com/KRM7/gapp.git --branch v1.0.0
sudo bash ./gapp/tools/install.sh

Once the library is installed, you can import it using find_package and then link against the gapp::gapp target provided by the library:

find_package(gapp REQUIRED)
target_link_libraries(YourTarget PRIVATE gapp::gapp)

Documentation

The documentation for the library can be found in the docs directory:

• Introduction
• Fitness functions
• Encodings
• Algorithms
• Genetic operators
• Stop conditions
• Metrics
• Miscellaneous

The API documentation is available here.

Examples

The examples directory contains several examples for using the library.

• Minimal example
• Single-objective optimization
• Multi-objective optimization
• Constrained optimization

---