A zone-based traffic assignment algorithm for scalable congestion reduction

Traffic assignment networks are networks with pseudo-static behavior: the network topology is constant but the cost of each edge changes in real-time. Extensive work has been completed in the literature to develop efficient traffic assignment algorithms in order to reduce traffic congestion. While some of these algorithms have proven to be effective, little attention has been paid to the matter of scalability in traffic networks. In this paper, we use zones to develop a hybrid approach to traffic assignment. We divide a traffic network into zones where the path within each zone is proactively stored, and paths between zones are reactively evaluated. This reduces the cost of route discovery. Using the Simulator of Urban MObility (SUMO), experiments were conducted to compare the zone-based system coined Z-BAR against a zone-free system. Between Z-BAR and a zone-free system, initial results showed Z-BAR introduces a speedup factor of up to 1.22