AI before GenAI: Ant Colony Optimization

Real ants find the shortest route between nest and food with no leader, no map, and no memory to speak of. Each ant follows two rules: wander toward the smell of pheromone, and drop pheromone as you walk. What’s interesting is that the intelligence isn’t in the ants - but in the trail!

My new post is a live demo, running in your browser. Sixty ants forage across six different maps (including the classic double-bridge lab experiment and a maze), a chart tracks how close they are to the true shortest path, and sliders let you break the colony: turn evaporation off and it can never forget a bad idea, turn it up and it can never commit to a good one.

The best part is dragging the food mid-run. The old trail fades, a new one condenses out of the traffic. Neat, right?

Read it and play with it here: https://www.kenreid.co.uk/blog/ant-colony-live.html