Schelling Segregation Model (2D)

Model Parameters

Tolerance (min. similar neighbors) 35%

Agent density 88%

Speed (steps / sec) 4

Run

Randomizing the grid — press Reshuffle any time.

Live Readout

Step

Happy agents

Segregation index (avg. similar-neighbor share)

How it works

Each cell on the grid is either empty, or home to an agent of type A or type B. Every step, an agent checks the fraction of its occupied neighbors (of its 8 surrounding cells) that share its own type. If that fraction is below the tolerance threshold, the agent is "unhappy" and relocates to a random empty cell.

Thomas Schelling showed in 1971 that even a very mild preference — wanting just a third of your neighbors to match your type — produces sharply segregated neighborhoods after only a few dozen steps, with no agent ever preferring segregation itself. The macro-level pattern emerges purely from many small, individually reasonable micro-decisions.

Higher tolerance → agents demand more similar neighbors → faster, starker segregation.
Lower tolerance → grid stays mostly mixed, few agents ever move.
Density controls how many cells start empty (empty cells are the only places agents can move to).

Type A resident Type B resident Empty cell