Schelling's segregation model is perhaps the best-known model
among social scientists, demographers and agent-based modelers. The
model shows that small preferences for one type of neighbor lead to
clusters of people of the same type. It also exemplifies how macro
behavior arises from micro behavior.
Observe (using initial conditions of 50-50) how a 20%
preference (desired of similarity) for both types lead to over 50%
resulting similarity of neighbors (see graph). Using a 30% preference,
similarity jumps to about 75% while 60% leads to almost 100%. Also
notice how clearer the clusters get on the simulation. However, notice
that 80% of preference does not lead to clusters. How would you
explain this outcome?
This model provides an explanation for segregated cities
(how segregated is your city).
Article: How
Segregated is Your City?