Sugarscape Wealth Simulator

What you just saw

Agents start with identical rules but varied vision and metabolism. After a few hundred steps, a small fraction of the population controls most of the wealth. The Gini coefficient climbs past 0.7 — inequality we normally associate with real economies — purely from heterogeneity and local harvest rules, with no policy, no coercion, no central planner.

The science behind it

Joshua Epstein and Robert Axtell introduced Sugarscape in their 1996 book Growing Artificial Societies. It is the canonical demonstration that significant economic inequality can emerge from simple individual-level rules. Extensions include inheritance, trade, combat, and cultural transmission — each one producing qualitatively different societies. It is a foundational model for computational economics.

Try these experiments

1. 1. Slow regrowth

   Settings: regrowth_rate=0.3

   What to look for: High inequality, small population. The landscape can't support many agents.

2. 2. Abundance

   Settings: regrowth_rate=3

   What to look for: Lower Gini, larger stable population. Abundance reduces inequality.

3. 3. Crowded start

   Settings: initial_population=800

   What to look for: Rapid die-off as agents exhaust the sugar faster than it regrows. Sharp population crash.

from scirouter import SciRouter
client = SciRouter(api_key="sk-sci-...")
result = client.simulation.run(
    model="sugarscape",
    params={"grid_size": 200, "initial_population": 50_000,
            "regrowth_rate": 1.0},
    steps=1000, seed=42,
)
print(result.metrics["gini_coefficient"])