mimic-world 0.1.0

Multi-company scenario engine for supply chain cascade simulation

mimic-world

Multi-company scenario engine for supply chain cascade simulation.

What happens to an entire system when a shock hits?

mimic-world simulates many companies responding to one event, watching them react to each other, and tracking the cascade. This is where supply chains come alive. This is where second-order effects emerge. This is where a port closure becomes an inflation event.

from mimic_world import World, Scenario, Twin

world = World()
world.add_twin(Twin.from_ticker("AAPL"))
world.add_twin(Twin.from_ticker("WMT"))
world.add_twin(Twin.from_ticker("FDX"))
world.add_twin(Twin.from_ticker("TSMC"))

world.connect("TSMC", "AAPL", relationship="supplier", weight=0.95, commodity="chips")
world.connect("XOM",  "FDX",  relationship="supplier", weight=0.40, commodity="jet_fuel")

scenario = Scenario.from_library("taiwan_strait_closure_30d")
result   = world.run(scenario)

result.print_summary()
print(result.cascade_timeline)
print(result.second_order_effects)
print(result.system_stability)   # "stabilizing" | "escalating" | "bifurcating"

How It Works

1. A Scenario defines initial shocks (semiconductor_supply: -0.65) and cascade rules
2. A World holds company Twins connected by a RelationshipGraph
3. The CascadeEngine runs time steps: each twin's LLM-powered decisions update a shared world_state
4. Connected twins see those updates and react — creating emergent second-order behavior

The 50-Scenario Library

Pre-built scenarios across 6 categories:

Category	Count	Examples
Geopolitical	10	Taiwan Strait closure, Russia energy cutoff
Supply Chain	10	Suez Canal closure, Shanghai lockdown
Macro/Financial	10	Fed hikes 200bps, Credit crunch 2008-style
Climate/Natural	10	Gulf Coast Cat 5 hurricane, Japan earthquake 9.0
Pandemic/Health	5	COVID-severity pandemic, Antibiotic resistance
Technology/Cyber	5	Global cloud outage, Critical infrastructure attack

# List all 50 scenarios
for s in Scenario.list_library():
    print(f"{s['id']:<45} severity={s['severity']:.0%}")

Install

pip install mimic-world

Ecosystem

mimic          → single company twin
mimic-bench    → grades predictions
mimic-forecast → real quantitative forecasts
mimic-world    → multi-company system + scenarios  ← HERE
mimic-sim      → 10,000-run Monte Carlo over worlds
mimic-signal   → real-time event detection

License

MIT