Traffic simulation of an urban network system using agent-based modeling

摘要

The problem of traffic congestion is attracting lots of attention with the aim of investigating efficient solutions. To achieve models that comprehensively describe the system, useful data from the network is needed. A suggestion is to mount detectors in selected regions and study the system responses to various stimuli. This, however, might be costly and impractical in many cases. Traffic simulators are beneficial tools for providing a simulated environment of the network. The current study tries to develop a traffic simulator within NetLogo, an agent-based environment. Using this simulator, the effect of different light plans on average delay time (ADT) of vehicles is investigated. The results indicate that in a normal-sized network a better choice is the simultaneous change of lights. A long length is appropriate for low entering frequencies and an average length is more efficient for high entering frequencies. In geometrically bigger networks, lights are better to change every other one with an average length for low entering frequencies and a long length for high entering frequencies.