Library Optimizer

A comprehensive Python library for metaheuristic optimization algorithms, featuring 20+ state-of-the-art optimization algorithms with both single-objective and multi-objective capabilities, all with a unified interface.

And especially there is a library for MATLAB: Libraries Optimizer for MATLAB

Features

• 20+ Optimization Algorithms: Grey Wolf Optimizer, Particle Swarm Optimization, Artificial Bee Colony, Whale Optimization, and more
• Multi-Objective Support: Automatic generation of multi-objective versions from single-objective algorithms
• Unified Interface: Consistent API across all algorithms for both single and multi-objective optimization
• Visualization: Built-in progress tracking and convergence plotting
• Benchmark Functions: Ready-to-use test functions for evaluation (single and multi-objective)
• Extensible: Easy to add new algorithms following established patterns
• Type Hints: Full type annotations for better development experience
• Automatic Detection: System automatically detects objective function type and selects appropriate solver

Installation

git clone https://github.com/phn1712002/LibraryOptimizer
cd LibraryOptimizer && pip install -e . 

Quick Start

Single-Objective Optimization

import numpy as np
from LibraryOptimizer import create_solver

# Define objective function (Sphere function - single objective)
def sphere_function(x):
    return np.sum(x**2)

# Create optimizer instance
optimizer = create_solver(
    solver_name='GreyWolfOptimizer',
    objective_func=sphere_function,
    lb=-5.0,  # Lower bound
    ub=5.0,   # Upper bound
    dim=10,   # Problem dimension
    maximize=False  # Minimization problem
)

# Run optimization
history, best_solver = optimizer.solver(
    search_agents_no=50,  # Population size
    max_iter=100          # Maximum iterations
)

print(f"Best solution: {best_solver.position}")
print(f"Best fitness: {best_solver.fitness}")

Multi-Objective Optimization

import numpy as np
from LibraryOptimizer import create_solver

# Define multi-objective function (ZDT1 benchmark)
def zdt1_function(x):
    f1 = x[0]
    g = 1 + 9 * np.sum(x[1:]) / (len(x) - 1)
    h = 1 - np.sqrt(f1 / g)
    f2 = g * h
    return np.array([f1, f2])

# Create optimizer instance - system automatically detects multi-objective function
optimizer = create_solver(
    solver_name='GreyWolfOptimizer',  # Same name, auto-detects multi-objective
    objective_func=zdt1_function,
    lb=np.array([0.0, 0.0]),  # Lower bounds
    ub=np.array([1.0, 1.0]),  # Upper bounds
    dim=2,                    # Problem dimension
    archive_size=100          # Archive size for multi-objective optimization
)

# Run optimization
history_archive, final_archive = optimizer.solver(
    search_agents_no=100,     # Population size
    max_iter=200              # Maximum iterations
)

print(f"Found {len(final_archive)} non-dominated solutions")
print(f"First solution: {final_archive[0].position} -> {final_archive[0].multi_fitness}")

Available Algorithms

All algorithms are available in both single-objective and multi-objective versions:

• Grey Wolf Optimizer (GWO) / Multi-Objective GWO
• Whale Optimization Algorithm (WOA) / Multi-Objective WOA
• Particle Swarm Optimization (PSO) / Multi-Objective PSO
• Artificial Bee Colony (ABC) / Multi-Objective ABC
• Ant Colony Optimization (ACO) / Multi-Objective ACO
• Bat Algorithm / Multi-Objective Bat
• Artificial Ecosystem-based Optimization (AEO) / Multi-Objective AEO
• Cuckoo Search (CS) / Multi-Objective CS
• Dingo Optimization Algorithm (DOA) / Multi-Objective DOA
• Firefly Algorithm / Multi-Objective Firefly
• JAYA Algorithm / Multi-Objective JAYA
• Modified Social Group Optimization (MSGO) / Multi-Objective MSGO
• Moss Growth Optimization (MGO) / Multi-Objective MGO
• Shuffled Frog Leaping Algorithm (SFLA) / Multi-Objective SFLA
• Teaching-Learning-based Optimization (TLBO) / Multi-Objective TLBO
• Prairie Dogs Optimization (PDO) / Multi-Objective PDO
• Simulated Annealing (SA) / Multi-Objective SA
• Genetic Algorithm (GA) / Multi-Objective GA
• More ....

The system automatically selects the appropriate version based on your objective function type.

Documentation

Full documentation is available at GitHub Repository.

Contributing

We welcome contributions! Please see our Contributing Guidelines for details.

Adding New Algorithms

To add a new optimization algorithm, follow the template in rules/single-objective.md. If you need to implement multiple fitness, follow the template in rules/multi-objective.md. Or the system will automatically generate the multi-objective version of your algorithm using the automatic generation system described in rules/auto-generation-multi.md.

Example Test Script

The library includes a comprehensive test script (test/example_test.py) that allows you to test specific optimization algorithms with various benchmark functions:

# Show available algorithms
python test/example_test.py -list

# Test a specific algorithm (e.g., GreyWolfOptimizer)
python test/example_test.py -name GreyWolfOptimizer

# Test another algorithm (e.g., ParticleSwarmOptimizer)  
python test/example_test.py -name ParticleSwarmOptimizer

The test script runs the following comprehensive tests:

• Sphere Function: Tests basic minimization capabilities
• Rastrigin Function: Tests performance on a multimodal function
• Maximization: Tests algorithm's ability to maximize functions
• Multi-objective ZDT1: Tests multi-objective optimization with 2 objectives
• Multi-objective ZDT5: Tests multi-objective optimization with 3 objectives

Each test validates that the algorithm can find reasonable solutions and provides detailed pass/fail reporting.

License

MIT License - see LICENSE file for details.

Citation

If you use this library in your research, please consider citing:

@software{LibraryOptimizer,
  author = {phn1712002},
  title = {Library Optimizer: A Python Library for Metaheuristic Optimization with Multi-Objective Support},
  year = {2025},
  url = {https://github.com/phn1712002/LibraryOptimizer}
}

Contact

• Author: phn1712002
• Email: phn1712002@gmail.com
• GitHub: phn1712002

Acknowledgments

This library builds upon research in metaheuristic optimization and implements algorithms from various scientific publications.