GeneticAlgorithmPython

Why we included this project

For optimization problems where gradients or closed-form solutions fall short, a genetic algorithm can be a pragmatic fallback, and PyGAD is one of the more approachable ways to run one in Python. You supply your own fitness function, then the library handles parent selection, crossover, and mutation across a range of built-in variants, covering both single- and multi-objective cases without making you write the evolutionary plumbing by hand. It also pairs well with machine learning work, since it can evolve the weights or hyperparameters of a Keras or PyTorch model as the population being optimized. The core install stays small, relying mainly on NumPy and cloudpickle, while plotting and deep-learning support come as optional extras. Teams wanting a scriptable optimization routine rather than a heavyweight platform will find the API easy to drop into existing NumPy-based code.