Evolutionary algorithms, explained with moving pictures

What it does This is a collection of standalone Python scripts demonstrating classic evolutionary computation techniques: genetic algorithms, evolution strategies, NEAT, and microbial GA. Each script targets a toy problem—matching a target phrase, solving traveling salesman, pathfinding, or training neural nets via OpenAI-style distributed evolution strategies. The code is paired with Chinese video and text tutorials on the author’s site, 莫烦 Python.

The interesting bit The repo treats visualization as a first-class citizen. Every algorithm ships with an animated GIF showing the population evolving in real time, which turns abstract selection-mutation-crossover loops into something you can actually watch. It’s pedagogy through motion.

Key highlights

• Covers GA, ES, NES, NEAT, and microbial GA in self-contained .py files
• Animated GIFs for each example, generated by the scripts themselves
• Bridges to neural nets: NEAT for supervised and reinforcement learning, plus distributed ES
• Author has spun off a packaged version, MEvo, for reuse beyond tutorials
• ~1,200 stars suggests it has found an audience among Chinese-speaking learners

Caveats

• Documentation and commentary are in Chinese; English speakers get only the code
• The repo is tutorial scaffolding, not a maintained library—expect to copy and adapt rather than pip install

Verdict Worth a look if you learn by reading code and watching populations converge. Skip it if you need production-grade evolutionary computation or English-language docs.