Abstract
Urban problems caused by fast urbanization become more and more complex, bringing new challenges to urban planning and its education. Therefor, proposition of artificial cities is put forward in this paper, based on which artificial urban planning research can be carried out. By adopting multi-agent system (MAS) method and according to features of Swarm simulation, the spatial evolution system model of an artificial city is built to study general process of urbanization, focusing on physical space, family agents, enterprise agents and interactions between (among) them in the city. Simulation of the model not only reproduces the process of a monocentric city's formation, but discovers that in a double-centered city's construction, administrative interference should be combined with the city's self-organizing evolution at a proper time. Simulation experiments and relative analysis show that the model is reasonable, which indicates artificial urban planning research is feasible and will offer effective support to urban planning education.