metaheuristics-lab
A desktop lab for watching metaheuristics think: genetic algorithms, ant colonies, swarms and more.
- OPENED
- MAY 2026
- COURSE
- Metaheuristic Algorithms
- ALGORITHMS
- 6
- SOURCE
- Public on GitHub
ROUTE · L8 Classroom · 6 stations
- DEC 2025 · Heritage station Meteo ESP32
- APR 2026 · Route complete TRL Backtesting
- MAY 2026 · Extension underway SistemaIVR
- MAY 2026 · Heritage station botAsesorCarreras
- YOU ARE HERE metaheuristics-lab
- SEP 2026 · Under construction AmortizaCore
Why it exists
You pick a problem type (continuous, binary, categorical, the traveling salesman or your own) and an algorithm, set its parameters in forms that adapt to the pair, and watch the population move live. At the end you get convergence curves, snapshots at each third of the run and an exportable GIF.
It was the course's final project, and like TRL and SistemaIVR, it was in the grave until Cursor revived it. I got to implement the imperialist competitive algorithm, based on colonization and empires, and I loved it because I love history.
A colony over Guadalajara
A real ant colony runs below. The ants leave Km 0 for the stations on the map, over Guadalajara's actual terrain. On the way back they lay pheromone: paths used often get stronger and the rest evaporate. Slopes slow them down, so the routes bend around the canyon and the hills.
It's written in Rust and runs in your browser as WebAssembly. Move your cursor over it to get in their way.