thieu1995/mealpy

A Collection Of The State-of-the-art Metaheuristic Algorithms In Python (Metaheuristic/Optimizer/Nature-inspired/Biology)

What it solves

MEALPY provides a centralized library for solving complex optimization problems using a vast array of meta-heuristic algorithms. It eliminates the need for researchers and developers to implement these complex mathematical optimizers from scratch, allowing them to quickly test, analyze, and compare different global search and black-box optimization methods.

How it works

The library implements population-based metaheuristics (PBMs), which are algorithms that use a group of candidate solutions to iteratively search for the best possible answer. It supports a wide variety of decision variables—including float, integer, binary, boolean, categorical, and permutation variables—to handle different types of problem spaces. The algorithms are categorized by their inspiration, such as evolutionary, swarm-based, physics-based, human-based, and biology-based methods.

Who it’s for

It is primarily designed for researchers and developers in the fields of AI, machine learning, and mathematical optimization who need to perform hyper-parameter tuning, feature selection, or solve general continuous and combinatorial optimization problems.

Highlights

  • Extensive Algorithm Collection: Includes 233 algorithms, covering original, hybrid, and variant versions.
  • Diverse Variable Support: Handles multiple data types including FloatVar, IntegerVar, StringVar, BoolVar, and SequenceVar for flexible problem definition.
  • Comprehensive Analysis Tools: Built-in capabilities to analyze convergence rates, scalability, robustness, and parameter performance.
  • Flexible Data Export: Supports saving results in multiple formats including CSV, JSON, pickle, PNG, PDF, and JPEG.

Related

  • Project
    Valdecy/pyMetaheuristic

    A Python library providing over 400 metaheuristic optimization algorithms with support for cooperative multi-island search, constrained optimization, and scientific benchmarking tools.

  • Project
    deanpeters/Product-Manager-Skills

    A library of 77 professional product management frameworks and structured skills designed to give AI agents and human PMs a shared foundation for high-quality, non-generic deliverables.

  • Project
    SciML/BlackBoxOptim.jl

    A global optimization package for Julia that uses stochastic and meta-heuristic algorithms to optimize non-differentiable functions.

  • Project
    SciML/Evolutionary.jl

    A Julia package providing a comprehensive suite of of evolutionary and genetic algorithms for complex optimization problems.

  • Project
    facebookresearch/Pearl

    A production-ready Reinforcement Learning library from Meta for building modular AI agents capable of handling complex, real-world environments with sparse feedback and dynamic action spaces.