Hierarchical Structure And Simulation Of Congestion Control For Urban Traffic Networks

In intelligent transportation systems, control strategies based on modelscan effectively solve the problem of optimal control for traffic networks,which is widely studied these years. However, when the scale of urban trafficnetwork is very large, this kind of control strategies will bring highcomputational complexity, and difficult to apply actually. This paper presentsa hierarchical structure for large scale urban traffic networks, reducing theonline computational complexity by decomposing the task. Meanwhile,considering the outstanding problems of traffic congestion in many Chinesecities, we propose a new problem, whose goal is to solve the congestionproblem, not to find the solution for one optimization taking some asperformance index. This paper provides control strategies and algorithms tosolve congestion. We use CORSIM to simulate real urban traffic, and thesimulation results show the effect of hierarchical structure and congestioncontrol. The main work of this paper is as follows:1) The interface of simulation based on CORSIM is built. The secondarydevelopment is realized using the C++program produced by CORSIM.Real-time data as well as computation result is passed between CORSIM andMATLAB program by RTE interface and MATLAB engine, laying thefoundation for the following research about control strategy.2) A hierarchical structure for large scale urban traffic networks isprovided, which divides a large scale network to several small subnetworks, and the structure contains two levels. The controller in the upper level gets theoptimized flow rate of the whole network by coordinating the flows betweensubnetworks. The model predictive controllers in the lower level try to followwhat the upper controller sets for them and find the best control signals withthe real-time traffic information, and then use the optimized signals at eachnode. The simulation results show that the hierarchical control is better thandecentralized control and closed to that of centralized control with muchlower computational complexity.3) Congestion control for urban traffic networks is presented. Thecorresponding model based on characteristics of saturated urban trafficnetwork is got, and then the standardized formulation of congestion control.The concept of congestion is defined. After these, we propose the sufficientconditions for judgment of redundant constraints in congestion controlproblem. The simulation results show the effect of congestion control.4) Saturated urban traffic network model is improved, and a new modelwhich can be suitable for all kinds of urban traffic network is proposed.Considering typically several fixed green ratio schemes are used in urbantraffic, a new congestion optimization problem is presented, which choose thescheme to meet the requirements.