IMAGES: Intelligent multi-agent system for freeway traffic flow simulation

Traffic simulators are very cost-effective and time-effective tools for improving infrastructure. Nowadays, microscopic model approaches have been in the limelight for traffic researchers, because they allow for more precise descriptions of the physical road environment as computer power continues to increase. The microscopic simulators focus on individuals within the system and, not the overall extent of the flow. In this paper, we describe IMAGES, which is an intelligent multi-agent system based on microscopic model approaches using longitudinal and lateral models. The multi-agent concept is sufficient to represent drivers' behaviour and cars' features. IMAGES is able to use scenario-based, event-driven, and real-time-based traffic data. Through such a simulation, we estimated freeway traffic flows excluding and including on/off ramps, and we confirmed that unexpected incidents cause traffic congestion. However, data gathered from sensors (loop detectors, cameras, RFID, etc.) should be used much more. This paper is a foundation stone to be built upon; after generating or acquiring data via IMAGES, the next step is studying on the data stream management system (DSMS), which deals with huge amounts of data in real-time.