Emergent phenomenon and the local information based DTA model

This paper proposes a simple but efficient dynamic traffic assignment (DTA) model based on real-time local-feedback information, which uses only the information acquired through sensors beneath the lanes to decide next turn to take with neither the confinement of route enumeration and O-D detail not the hunting phenomenon caused by exploiting global information. Some simulation-based experiments and comparisons with other DTA methods are made to evaluate the system performance of the local information based technique both in ordinary congested urban traffic conditions and in unusual incident occasions. Finally, robustness and effectiveness of the model is explained as a emergent phenomenon by comparing the local information based model in traffic network with the local feedback decentralized planning in ant colony.