inZORi Social Density Effect | Emergent Collective Behavior under Density Regimes

Abstract

A 2D world with autonomous inZORi organisms is run under three population density regimes — LOW (20 initial organisms), MEDIUM (60), and HIGH (140) — across 24 seeds × 800 steps each. The study measures how density affects emergent collective behaviors: mortality, spatial distribution, clustering, genome diversity, and energy stability. No social rules are programmed — all effects emerge from organism interactions through shared environmental resources. Statistical analysis (t-tests, 95% CI) confirms significant density effects on mortality (p<0.0001), spatial spread (p=0.0002), and clustering (p=0.003).

Seeds per density regime

800

Steps per run

+87%

Mortality increase: LOW→HIGH

+13.7%

Spatial spread increase: LOW→HIGH

−22.9%

Clustering decrease: LOW→HIGH

Stable

Mean energy across all densities (~4.0–4.4)

1. Experimental Setup

Density Regimes

LOW density — 20 initial

Final pop. (mean)386.5

Mortality rate0.18%

Mean distance37.6

Clustering3.71

Genome diversity0.222

Energy mean3.95

MEDIUM density — 60 initial

Final pop. (mean)387.3

Mortality rate0.24%

Mean distance42.2

Clustering3.11

Genome diversity0.221

Energy mean4.08

HIGH density — 140 initial

Final pop. (mean)391.7

Mortality rate0.34%

Mean distance42.8

Clustering2.86

Genome diversity0.211

Energy mean4.39

What is Being Measured

Mortality rate: fraction of organisms that die per step (energy depletion)
Mean distance: average distance between organisms — measures spatial spread
Clustering: mean number of neighbors within a local radius — measures cohesion
Genome diversity: variance in genome values across population — measures evolutionary diversity
Energy mean: average stored energy per organism — measures resource health

2. Results

Fig 1 — Aggregated metrics comparison across LOW / MEDIUM / HIGH density regimes (24 seeds each, error bars = 95% CI). Mortality increases monotonically; mean distance increases; clustering decreases. Energy mean remains stable across all regimes.

Fig 2a — Population dynamics: LOW density (initial 20 organisms, 24 seeds). Final mean: 386.5. Rapid initial growth followed by stabilization.

Fig 2b — Population dynamics: MEDIUM density (initial 60 organisms, 24 seeds). Final mean: 387.3. Growth rate slower; convergence occurs earlier.

Fig 2c — Population dynamics: HIGH density (initial 140 organisms, 24 seeds). Final mean: 391.7. Higher initial competition visible in early mortality spike; convergence to similar final population as other regimes.

Live Visualization

Real-time organism movement under each density regime (GIF animations, 24-seed mean behavior):

LOW density (20 initial). Sparse initial coverage; high individual freedom; slower clustering.

MEDIUM density (60 initial). Balanced coverage; intermediate clustering patterns visible.

HIGH density (140 initial). Intense early crowding; organisms spread out faster to reduce competition; reduced cohesion.

Statistical Results

Metric	LOW (95% CI)	HIGH (95% CI)	t-stat	p-value	Significance
Mortality rate	0.00181 [0.00173, 0.00189]	0.00338 [0.00325, 0.00351]	−20.5	<0.0001	p<0.0001
Mean distance	37.6 [35.5, 39.7]	42.8 [41.1, 44.5]	−3.76	0.00017	p=0.0002
Clustering	3.71 [3.28, 4.13]	2.86 [2.51, 3.21]	3.02	0.0025	p=0.003
Genome diversity	0.222 [0.219, 0.224]	0.211 [0.210, 0.212]	7.37	<0.0001	p<0.0001
Energy mean	3.95	4.39	—	—	Not sig. (stable)

3. Key Findings

      Mortality +87% (LOW → HIGH): crowding creates resource competition that the genome cannot fully compensate through behavioral adaptation alone.
Mean distance +13.7%: at high density, organisms spontaneously spread out — an emergent avoidance behavior reducing local resource competition.
Clustering −22.9%: less spatial cohesion at high density — crowding pressure overrides the natural tendency to cluster near resource-rich zones.
Genome diversity decreases at high density (0.222 → 0.211, p<0.0001): stronger selection pressure under competition reduces genome variation — selection is more intense when resources are scarce per capita.
Energy remains stable despite density differences: organisms compensate at the individual level through behavioral adjustments, even if population-level mortality increases.
Critical threshold between MEDIUM and HIGH: the mortality slope steepens significantly at the HIGH regime transition, suggesting a nonlinear density effect above ~60 organisms.
Final population converges (~386–392 across all regimes): regardless of initial density, inZORi populations reach similar carrying capacity by step 800.

    

4. What This Demonstrates

Density-Dependent Emergent Behavior

This study demonstrates that inZORi populations exhibit density-dependent emergent behaviors analogous to those observed in biological ecosystems: increased mortality, spatial dispersal, reduced clustering, and narrowed genome diversity under high population pressure. None of these responses were programmed — they emerge from the interaction between organism survival drives and shared environmental resource constraints.

The finding that genome diversity decreases under high density is particularly significant: it mirrors the evolutionary dynamics of bottleneck populations, where intense competition acts as a strong filter on genetic variation. This is the same mechanism responsible for reduced biodiversity in overcrowded natural habitats.

The stable energy mean across densities suggests that individual organisms successfully adapt their foraging strategy to their local conditions, even when population-level mortality increases. This decoupling between individual adaptation and population-level pressure is a hallmark of robust emergent systems.

Applications: Urban planning (resource allocation under population density), ecological modeling, distributed robotic swarms, epidemiology (density-dependent disease transmission analogs).

5. Reproducibility

Framework: inZORi v1.0 | Domain: Social dynamics / emergent collective behavior

Runs: 24 seeds × 3 density regimes × 800 steps = 72 total runs

Density regimes: LOW (initial 20), MEDIUM (initial 60), HIGH (initial 140)

Statistical tests: Two-sample t-tests (Welch) with 95% CI; all reported p-values are two-tailed

Metrics: Mortality rate, mean inter-organism distance, clustering coefficient, genome diversity (variance), energy mean

Note: inZORi genome structure and selection mechanics are proprietary. Environmental setup (2D continuous world, shared resources) and all reported metrics are fully disclosed.