Mesoscopic Traffic Flow Model for Agent-Based Simulation

Traffic simulation is a key element of agent based models as the various agents decisions impact and are impacted by the realized travel times and delays in the traffic network. Here we present the mesoscopic traffic flow model implemented in POLARIS transportation systems simulator. The model is mesoscopic in an attempt to obtain enough for applications that requires microscopic data while still inheriting the computational efficiency of macroscopic traffic flow models. The model is macroscopic at the link-level with dynamics based on the Newells Model and microscopic at the node level. We present link-level and network-level results for one mid-sized traffic network. The results show that the model can combine accuracy, level of details and computational efficiency.