AI Overseer: Resource Allocation Across a Simulated World

AI Overseer Policy

Resource pool growth 8

AI intervention strength 85%

Simulation speed 2.0 ticks/s

World State

Resource pool: 20.0 units

Total population

Avg. health

Inequality (Gini)

0.00

Ticks elapsed

How it works

28 settlements sit on a simulated world. Each tick, settlement i with population p_i and fixed productivity q_i produces and consumes:

output_i = BASE_PROD · q_i · p_i
need_i   = CONSUMPTION · p_i
balance_i = output_i − need_i   (surplus if >0, deficit if <0)

Positive balances are taxed into a shared pool; the pool also grows from the "Resource pool growth" slider. The AI overseer then redistributes the pool by one governing policy:

Utilitarian:    fill largest-population deficits first, until the pool runs out
Egalitarian:    split the pool evenly across all settlements in deficit
Need-Weighted:  alloc_i = pool · deficit_i / Σ deficit_j
Laissez-Faire:  alloc_i = 0  (no redistribution — baseline)

Only a fraction of the computed allocation — the "intervention strength" — is actually delivered, modelling a more or less hands-on deity. Each settlement's health then moves toward its net balance (local + AI-delivered), and population grows when health is high, shrinks when it is low. The inequality readout is the Gini coefficient of population across all 28 settlements:

G = (2·Σ i·x_i(sorted) − (n+1)·Σ x_i) / (n·Σ x_i),  0 = perfectly equal, 1 = maximally unequal

This is the same class of problem a real allocation algorithm solves — a scheduler dividing CPU time, a grid balancing power, or aid distributed across regions — with the policy choice trading off total output against fairness.

Thriving settlement Struggling Starving Resource transfer

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.