An agent-oriented software platform for urban traffic networks modeling, analysis and control

The paper presents a software application for modeling, simulation and analysis of urban traffic networks behavior in different functioning contexts. A module of control techniques, based on different approaches (as genetic algorithms, fuzzy and neural networks) allows the platform to provide decision support for context-based traffic control in large urban areas. The application is based on a three-layered control architecture: the basic level monitoring and controlling junctions, the middle one ensuring smooth functioning of sub-networks, by the coordination of interconnected junctions and the top level ensuring a globally optimized functioning of the system.