New release of Gradient-Free-Optimizers with two new evolutionary algorithms

The GitHub repository Gradient-Free-Optimizers offers a collection of optimization techniques that do not depend on gradients, suitable for numerical discrete search spaces. It features a user-friendly API for defining objective functions and search spaces, and incorporates modern optimization methods like Bayesian optimization, which are particularly effective for expensive objective functions. The library is rigorously tested, boasting over 400 tests to validate the performance of its algorithms.

The repository supports various optimization algorithms categorized into four main types: local optimization methods such as Hill Climbing and Simulated Annealing; global optimization techniques including Random Search and Grid Search; population-based methods like Particle Swarm Optimization and Genetic Algorithms; and sequential model-based optimization strategies such as Bayesian Optimization and the DIRECT algorithm.

Installation of the package is straightforward via pip, using the command "pip install gradient-free-optimizers." Additionally, the repository provides examples demonstrating the optimization of both convex and non-convex functions, as well as the optimization of machine learning hyperparameters. For further details, users can access the official documentation linked within the repository.