Tokyo Subway Model: The Evolutionary Math Behind the Perfect Transit Map
How a brainless slime mold redesigned the Tokyo railway system in just 24 hours, proving that biological engineering is often more efficient than human design.
Tokyo Subway Model: The Evolutionary Math Behind the Perfect Transit Map
In 2010, a team of researchers led by Toshiyuki Nakagaki at Hokkaido University performed an experiment that shocked the world of urban planning. They took the map of Tokyo and its surrounding cities, placed oat flakes on the locations of major railway stations, and released a single slime mold (Physarum polycephalum) at the center (the location of Tokyo’s main hub).
What happened next became a landmark study in biomimicry. Within just 24 hours, the brainless blob had grown a network of veins that almost perfectly replicated the existing Tokyo subway system. In some areas, the blob’s design was actually superior to the one built by human engineers.
Biological vs. Human Engineering
Human engineers spend decades—and billions of dollars—optimizing transit networks for efficiency, cost, and reliability. They use supercomputers to calculate the shortest paths between thousands of hubs.
The slime mold, however, does this using nothing but hydrodynamic pressure.
- Foraging Strategy: The blob initially spreads out in all directions to find food.
- Tube Reinforcement: Once it finds multiple food sources (stations), it begins to strengthen the veins that are high-traffic and retract those that are redundant.
- The Result: A network that balances efficiency (short paths) with resilience (redundancy in case a track breaks).
The Speed of “Thought”
While cities take decades to build, the blob works at a relentless, biological pace.
- Standard Speed: The blob typically moves at about 1 centimeter per hour.
- The Hunger Factor: If the blob is hungry and detects a nearby food source, it can accelerate its internal shuttle streaming, reaching speeds of up to 4 centimeters per hour.
In the Tokyo experiment, this meant the entire complexity of one of the world’s most sophisticated rail systems was “re-invented” in less than a day.
Why It Matters: Beyond Railways
The Tokyo Subway Experiment isn’t just about trains. It is a proof of concept for Decentralized Optimization.
- Fault Tolerance: If a vein in a slime mold network is cut, the flow simply reroutes through an existing alternate path. This is a blueprint for designing more resilient power grids and internet routers.
- Energy Efficiency: The blob finds the most energy-efficient configuration for transporting nutrients without a central “brain” making the decision. This could lead to better logistics algorithms for shipping and delivery networks.
The next time you are sitting on a train, remember that a brainless bag of yellow goo underneath a log in the forest might have designed a better route than the engineers who built the tracks.
Fascinated by slime mold math? Explore our Slime Algorithms Guide to see how code mimics the blob.
Origin and E-E-A-T
- Source: Le Monde: “Pourquoi le blob fascine les scientifiques.”
- Research Paper: Science magazine (2010), “Rules for Biologically Inspired Adaptive Network Design.”
- Key Scientists: Toshiyuki Nakagaki (Hokkaido University), Atsushi Tero.
Sources, Review, and Trust Signals
Origin Of Information
Le Monde: 'Pourquoi le blob fascine les scientifiques' & the Toshiyuki Nakagaki study (2010). (https://www.lemonde.fr/)
Editorial Review
Status: in review
Reviewed by: Slime Mold Club Editorial Team
Last reviewed: 2026-02-11
Concepts Used
Related Guides
SMT Analysis: How the Steiner Minimum Tree Rules Slime Mold Design
same pillar
The 720 Sexes: Decoding the Complex Mating Game of the Blob
same pillar
The 720 Sexes: A Masterclass in Dating
same pillar
Genetic Blueprints: The Code of the Syncytium
same pillar
Arterial Navigators: The Future of Medical Goo in Human Veins
same pillar
Between States: The Fluid-Solid Boundary of the Slime Mold Body
same pillar
Curious for more?
Your blob is always growing. Check out these related guides to keep her happy.