基于模型预测的宏观交通路网优化研究

Research on Optimization of Macroscopic Traffic Network Based on Model Predictive Control

为了缓解日益严峻的交通拥堵,解决大尺度交通网络微观建模复杂度高以及路网整体性能优化等问题,设计了基于模型预测的宏观交通网络优化模型。先获取大尺度交通网络的宏观交通流图(MFD),并在此基础上设计了模型预测控制器(MPC),改进了道路交通模型,以满足模型预测控制对预测模型的准确性要求。设计系统评价函数,在优化总时间花费(TTS)的同时,控制路网累计车辆总数逼近标准值,从而提高路网运行效率与流畅度。实验结果表明,在该模型作用下,车辆的总体时间花费与平均流速都得到了优化。

In order to alleviate the problem of increasingly severe traffic congestion,the high complexity of micro-modeling of large-scale traffic network and the optimization of the overall performance of road network,the model of macro-traffic network optimization based on model prediction is designed.Firstly,the large-scale transport network Macroscopic Fundamental Diagram(MFD)is obtained.Next,MPC(Model Predictive Control)controller is designed on this basis,and the road traffic model is improved to meet the accuracy requirements of the model predictive control.Meanwhile,the evaluation function of the system is designed,the total time spent(TTS)is optimized while the total number of vehicles on the road network to approach the standard value is controlled to improving the road network operation efficiency and fluency.Finally,the experimental results show that the total time spent and the average velocity of the vehicle are optimized by the model.