Urban traffic simulated from the dual representation: Flow, crisis and congestion

We propose a traffic simulation model for urban system based on the dual graph representation of a urban road network and with a random entering vehicle rate. To avoid the shortcoming of "Space Syntax" of ignoring the road's metric distance, we consider both the motion of the vehicles along roads and the navigation of the vehicles in the network. Simulations have shown some basic properties of urban traffic system, such as flux fluctuation, crisis and dissipation, phase transition from a free flow to jams, overall capacity, the distribution of traveling time, and the fundamental diagram. The system's behavior greatly depends on the topology of the transportation network. A well-planned lattice grid can keep more vehicles travelling. The critical entering vehicle rate is much greater in lattice grid than in a self-organized network. The vehicles have to travel longer time in a self-organized urban system due to the navigation cost.