Procedurally generated cities through multiple time periods

Abstract

This thesis introduces a method for the automatic generation of cities that show their development over time through changes in the city layout. Our method is based on an existing agent-based city generation and simulation method which uses its agents to scour the terrain and place roads and buildings on patches of land that they consider valuable. By introducing time periods that can change the agent’s behavior over time, the city layout also changes over time and as a result the city gets a unique pattern that reflects its development. In order to gain an understanding of the behavior of our model, we run a variety of simulations with different parameter settings to study the parameter influence on the agents, and thereby the model itself.