Research On Optimal Control Of Regional Traffic Coordination Based On Dynamic Road Network Division

In today’s society,the problem of traffic congestion is becoming more and more serious with the rapid development of the economy and the acceleration of the process of urbanization.Therefore,the research on the coordinated control of intelligent traffic signals has become an important way to optimize the operation efficiency of the traffic network.Because the structure of urban traffic road network and the traffic flow in the road network are often very complicated,this article proposes a real-time and reasonable dynamic division method based on the divided control sub-areas according to the insufficiency of the existing regional traffic road network sub-areas division method.An improved signal timing strategy is established.Finally,in order to optimize the coordinated control of different sub-intervals,a two-level planning model for the overall coordination of the regional traffic network is constructed to realize the overall coordinated operation of the road network.The main research contents are as follows:Firstly,in view of the high complexity of the regional road network and the shortcomings of the existing sub-zone dynamic partitioning methods,with the goal of optimizing the coordinated control of regional traffic,a dynamic regional road network partitioning method that combines improved correlation and Newman’s algorithm is proposed.First,it comprehensively analyzes the two factors of fleet dispersion and traffic density,and proposes an improved relevance model;then,using modularity Q as the division standard,the intersection relevance is used as the edge weight,and the traditional unauthorized community agglomeration algorithm is improved to make It can dynamically divide the road network into different sub-areas according to the characteristics of traffic flow.The simulation experiment results show that the proposed improved division method can effectively combine the characteristics of actual traffic flow to make more accurate real-time dynamic division of road network sub-areas.Secondly,in view of the dynamic characteristics of traffic flow in the divided sub-area road network and the deficiencies of existing delay models,an improved signal timing strategy for intersections is proposed.First,the existing vehicle average delay model is improved.By analyzing the influence of different signal phases at downstream intersections on the flow of outgoing traffic at upstream intersections and the influence of queued vehicles on the phase difference at the end of the phase,it is further deduced that downstream intersections are in different traffic Vehicle delay calculation method under flow conditions;secondly,the graph convolutional neural network is used to predict the queuing flow accumulated in different phases of each entrance lane in the downstream intersection,which effectively predicts the trend of queued traffic flow at the end of different road phases,and then Establish an intersection signal timing strategy based on the queuing traffic flow prediction delay model,and use genetic algorithm to optimize the solution;finally select a divided sub-area for simulation test,the experimental results prove that the GCN prediction model used has good The proposed signal timing strategy can give the optimal timing plan according to different traffic flow conditions,which effectively improves the traffic operation efficiency of the sub-area.Finally,for the coordinated control of the divided sub-intervals,the maximum capacity of each sub-area in the regional road network is used as the upper function optimization objective,and the average vehicle delay of each entrance lane in the regional road network is minimized as the lower function optimization The goal is to establish an overall coordinated two-level programming model for the regional traffic network,and use the genetic simulated annealing algorithm to optimize the solution.In this two-level programming model,the upper model optimizes the optimal public cycle duration of each sub-area,and the lower model optimizes the optimal green light duration of each intersection in the regional road network and the phase difference between adjacent intersections.The actual regional traffic network is selected for simulation comparison.The experiment proves that the regional traffic coordination optimization control model based on dynamic road network division proposed in this paper can effectively reduce the average delay time of regional traffic flow,improve the efficiency of traffic network operation,and alleviate traffic congestion.