An agent based approach for modeling traffic flow

Driver behavior is a key factor that gives rise to traffic congestion. In this paper we make use of agent based modeling (ABM) in simulating a traffic system. To investigate how drivers´ behavior effect on the performance of the traffic system, we modeled four different types of car drivers. In order to enhance the traffic system´s performance, we used a genetic algorithm to optimize the schedule of traffic lights to maximize the overall network throughput. Moreover, we built our model using a 3D modeling environment (Blender) for a more realistic simulation of vehicles´ motion.