| In recent years,new ideas have been developed in the study of path optimization with the integration of intelligent transportation technology.In view of the local traffic congestion caused by the rapid increase of motor vehicle ownership,this dissertation chooses the road around Xizhanshizi in Qilihe District,Lanzhou City,Gansu Province,as the research target area to carry out the research on the optimization of traffic flow in unbalanced road network.The main contents are as follows:(1)In order to further verify the necessity of urban traffic flow planning problem,in the introduction of some basic traffic flow running state evaluation index and some regularity on the basis of urban traffic flow characteristics,In this dissertation,the principle of game theory is added to the model of multi vehicle movement in urban traffic,by building a game matrix,analyzes the contradiction between local optimum and global optimum in path planning,The limitation of one-sided pursuit of the optimal scheme of single vehicle path is revealed,and the nonlinearity and complexity of urban traffic problems are confirmed.(2)In order to resolve the problem of collecting and matching dynamic traffic flow data,the data characteristics of road network topology,congestion level and instantaneous traffic flow are respectively studied.In the absence of perfect monitoring technology and monitoring equipment,the traffic data provided by network map service providers are used for field observation of the traffic data in the target area to ensure data quality.By means of fuzzy clustering,the mapping relationship between instantaneous traffic flow and congestion level is summarized,which provides the data foundation for the case study.(3)In view of the problem of optimization of urban traffic flow in unbalanced road network,in this dissertation,the scheme of traffic flow distribution at each intersection of the road network is optimized as the research object,and the genetic algorithm is selected to optimize and resolve the special situation of different sections in the region to make a detailed analysis,construct the individual coding scheme,improve the algorithm process logic and objective function,and enhance the efficiency of the algorithm and the value of the results.The optimized road network congestion state is compared with the actual traffic congestion state to prove the effectiveness of the overall path optimization.(4)In the design of the fitness analysis module of the optimization algorithm,this dissertation takes the traffic congestion level of relevant sections as the state variable to make short-term traffic flow prediction.Combined with the actual situation of available data in the target research area,the Long Short-Term Memory neural network is selected as the prediction model to make short-term prediction of traffic flow congestion level and congestion trend,and an important module of fitness evaluation after path optimization is constructed.Through the research and analysis of the traffic data in the target area,the proposed scheme which combines the short-term traffic flow prediction method with genetic algorithm has a good effect on the traffic path optimization problem under the condition of unbalanced road network.It can provide new ideas and directions for the study of unbalanced urban path optimization. |
