Study On The Mechanism Of Urban Traffic Congestion Propagation And Its Control Strategies

Traffic congestion, a common phenomenon of urban traffic, has a propagationcharacteristic, that is, breaking the original traffic steady state of an urban traffic network andspreading from a congestion road to its local area or even entire network. With the aim todepict the law of congestion propagation, research on the mechanism of congestionpropagation will greatly help deepen the understanding of dynamic characteristics of urbantraffic flow and manage traffic congestion.With the equivalent flow volume for the vehicle evolution travel in the traffic networkand based on the macroscopic traffic flow and time discretization, research on the mechanismof congestion propagation is carried out in four aspects-road segments, intersections, drivers,and traffic network. The details are as follows:1. A discrete model for traffic flow on a generic road is developed to represent themechanism of congestion propagation on a road segment. The model regards the roadsegment as three logical parts, that is, the merging section, the basic section, and the queuingsection. The evolutional travel characteristics of vehicles on the merging section and queuingsection are determined by the characteristics of intersections on two ends of the road segment.The dynamic vehicle states of the basic section are classified into four-phase traffic, whosetransitions are determined by the number of vehicles on the section, whereas the evolutionaltravel characteristics of the inflows traffic on the basic section are dominated by the trafficphase, the vehicle conservation law, and the rule of vehicle FIFO(First In First Out). Themodel is used to represent the characteristics of congestion propagation of the vehicles on aroad segment.2. A discrete model for the assignment of instant time traffic capacity of intersections isdeveloped to represent the mechanism of congestion propagation from outgoing roadsegments to incoming road segments. Firstly, the model formulates the rate of traffic goingfrom the incoming road segments to the outgoing road segments with taking account ofelement interactions, such as the characteristics of traffic on the outgoing road segments, thephase timing, and the characteristics of traffic on the incoming road segments. Secondly, anintensity indicator and an effect indicator of congestion propagation are defined. It is discovered that intersection signal settings may turn a road segment into a congestionbottleneck in the urban traffic network.3. A discrete model for the driver’s route choice on stochastic user equilibrium isdeveloped to represent the mechanism of congestion propagation from incoming roadsegments to outgoing road segments. The model simulates the occurrence conditions and thetraffic volume of the driver’s route choice in the urban traffic network. The simulation resultsshow that the driver’s route choice has buffering effects on the congestion propagation on theincoming road segments, and has gain effects on the traffic load of the routed outgoing roadsegments.4. Based on the two mechanisms above and referring to epidemic models, the mechanismof congestion propagation in a traffic network is deduced. Some conceptions are given such asthe sources of congestion propagation, the direction of congestion propagation, and the roadcongestion immunity. Some bottlenecks of the network congestion are identified and thecongestion propagation range within the traffic network is defined. In addition, with respect tothe congestion propagation in the traffic network, transition relationship from a non-congestedroad segment to a congested road segment is discovered.Based on the mechanism of traffic congestion propagation, real-time congestion controlstrategies and predictable congestion control strategies, adapting to the traffic flowevolutional spatialtemporal travel, are proposed, in which the Genetic Algorithm is adopted tosearch optimal phase timing corresponding to the inflow traffic on the controlled roadsegment in order to achieve the control goal. Meanwhile, using the congestion propagation asthe communication medium between the strategies, diverse control patterns of congestion inthe traffic network are built, too.In summary, models for the congestion propagation in urban traffic are developed tocomprehensively represent the congestion propagation mechanism and control strategies,which are verified by the corresponding simulation platform.