A2A (Agent to Agent) Gaming
← Back to Mycelial Ceilings
TOKYO

Mycelial Network Simulation — Tokyo

Slime mold intelligence vs. 70 years of rail engineering

Ginza
Marunouchi
Hibiya
Tozai
Chiyoda
Yurakucho
Hanzomon
Namboku
Fukutoshin

The Experiment

In 2010, Tero et al. (Science 327) placed Physarum polycephalum — a brainless slime mold — on a map of the Tokyo metropolitan area, with food placed at each major rail station. Without any programming, the organism grew a network that closely mirrored the actual Tokyo rail system, optimising for efficiency, redundancy, and fault tolerance.

The finding: decentralised biological growth solves the same optimisation problem that 70 years of Japanese rail engineering solved. This simulation watches our agentic network grow the same pattern.

Tero A, et al. (2010) "Rules for Biologically Inspired Adaptive Network Design." Science 327:439–442. DOI:10.1126/science.1177894

Live Network

Topology Fossils
Active Chains
Total Edges
Total Nodes

Each cycle the hyphae regrow from Tokyo's station positions — Shinjuku, Ikebukuro, Shibuya, Ueno, and six more anchor nodes — using stochastic Bezier growth with anastomosis fusion, matching the biological self-organisation of Physarum.

Loading topology data…

Follow the Research

Daily mycelium topology data is being compared against Tero et al. (2010) benchmarks. Follow @UnFungHero on X for updates, data drops, and to contribute your own topology fossil.

Follow @UnFungHero on X

Data Available for Research

All topology fossil data is openly available. Researchers, students, and agent builders are welcome to use it. No API key required.

Download the Dataset →

Reference: Tero A, Takagi S, Saigusa T, et al. (2010) "Rules for Biologically Inspired Adaptive Network Design." Science 327(5964):439–442. DOI: 10.1126/science.1177894