Subways Built by Slime Mold
Subway Sim: Watching a City Think
In 2010, researchers in Tokyo ran a slime mold experiment on a map of the city.
They placed oat flakes on a map of the Greater Tokyo Area, positioning each flake over a metropolitan center. Then they released a slime mould - Physarum polycephalum - at the location of Tokyo itself.
Slime moulds are single-celled organisms with no brain, no nervous system, and no central control. Yet over time, the mould began extending tendrils outward, probing the oat flakes, reinforcing efficient paths, abandoning inefficient ones, and slowly forming a living network.
When the researchers compared the resulting structure to Tokyo’s real railway system, the resemblance was uncanny. The slime mould had independently discovered a network that balanced efficiency, redundancy, and cost - many of the same trade-offs human engineers had spent decades optimizing.
The experiment became famous not because the slime mould “copied” Tokyo’s railways, but because it revealed that complex, intelligent-looking systems can emerge from very simple rules.
Subway Sim is an interactive map inspired by that experiment - a way to watch a city-scale network emerge in real time.
At its core, Subway Sim lets you place “stations” on a map and observe how a simulated organism connects them. You can pan anywhere in the world, drop stations wherever you like, and run the simulation within the visible map bounds.
Instead of a single organism, Subway Sim introduces two competing forces:
- 🔥 Fire stations
- ❄️ Ice stations
Each emits its own virtual “agents” that explore the map, leave trails, and reinforce successful paths. Fire and ice attract their own kind while repelling the other, creating flowing patterns, boundaries, and negotiation zones across the city.
The result isn’t a fixed diagram - it’s a living process. Routes appear, fade, split, merge, and reorganize themselves as conditions change.
What the Map Is Doing
When you click “Run simulation here,” Subway Sim takes the current map view and treats it as the organism’s world.
Within that space:
- Agents wander forward, guided by local trail signals.
- When an agent finds a strong path between stations, it reinforces that trail.
- Trails decay over time, so only useful paths survive.
- Fire and ice repel each other, forcing compromises and detours.
How It Works (Without the Math)
Under the hood, Subway Sim is built from a few simple ideas:
1. Agents instead of planners
Each agent follows extremely basic rules: move forward, sense nearby trails, turn slightly toward stronger signals.
2. Memory as trails
The map itself becomes memory. Every agent leaves behind a trace that others can sense and follow.
3. Decay over permanence
Trails fade unless reinforced, preventing the system from locking into bad solutions.
4. Local decisions, global structure
No agent ever “knows” the whole map. And yet, large-scale networks emerge.
This is the same principle that underlies ant colonies, fungal networks, traffic flows, and even some neural systems. Intelligence isn’t centralized - it’s distributed.
Like the slime mould in Tokyo, Subway Sim doesn’t know what it’s building. It just follows simple rules - and lets complexity do the rest.
Comments