城市道路交通拥挤状态判别及预测研究

发布时间：2018-09-08 21:05

【摘要】：城市有限的道路资源难以承载交通量的快速增长,导致交通拥堵问题的出现,交通拥堵预测是解决交通拥堵问题的重要步骤之一。但由于影响交通系统的因素复杂繁多,且各种交通参数具有较强随机性和不确定性,使得交通拥堵预测研究难以开展,预测成功率及可靠性往往不高,针对这一问题,本文借鉴马尔可夫理论及灰色预测理论,构建了适用于交通拥堵预测的灰色GM(1,1)-加权马尔可夫预测模型,并将该模型应用于实例研究中。具体研究过程如下：首先,在回顾国内外研究现状的基础上,给出了拥堵的定义、分类、成因和特征。对经典的拥堵识别算法和常见的速度预测模型进行了分析；其次,探讨速度预测与拥堵识别的关系和基于速度的交通拥堵预测的原理,并确定相应的速度阈值标准,基于灰色预测理论,结合马尔可夫链预测原理,建立灰色GM(1,1)-马尔可夫预测模型用于交通拥堵预测,并在此基础上对该模型进行加权改进以获得更好的预测成功率；最后,将该模型应用于石家庄市主干路——建设大街的拥堵预测实例研究中,对该路段未来4天内6个不同时刻的拥堵状态进行了预测识别,并与灰色GM(1,1)预测模型、灰色GM(1,1)-马尔可夫预测模型的预测结果相比较。结果表明,该模型的识别成功率超过66%,优于灰色GM(1,1)预测模型和灰色GM(1,1)-马尔可夫预测模型,从而表明本文所建立的预测模型具有较好的识别准确率及可靠性。
[Abstract]:Limited urban road resources are difficult to support the rapid growth of traffic volume, leading to the emergence of traffic congestion problem, traffic congestion prediction is one of the important steps to solve the traffic congestion problem. However, due to the complexity of the factors affecting the traffic system and the strong randomness and uncertainty of various traffic parameters, it is difficult to carry out the research of traffic congestion prediction, and the success rate and reliability of traffic congestion prediction are not always high. Based on Markov theory and grey prediction theory, a grey GM (1) -weighted Markov forecasting model for traffic congestion prediction is constructed in this paper. The model is applied to a case study. The specific research process is as follows: firstly, the definition, classification, causes and characteristics of congestion are given on the basis of reviewing the current research situation at home and abroad. The classical congestion identification algorithms and common speed prediction models are analyzed. Secondly, the relationship between speed prediction and congestion identification and the principle of traffic congestion prediction based on speed are discussed, and the corresponding speed threshold standard is determined. Based on the grey prediction theory and the Markov chain prediction principle, the grey GM (1k-1) -Markov forecasting model is established for traffic congestion prediction, and the weight of the model is improved to obtain a better prediction success rate. Finally, the model is applied to the case study of traffic congestion prediction on the main road of Shijiazhuang City-Construction Street. The congestion state of this section at 6 different times in the next 4 days is forecasted and identified, and it is compared with the grey GM (1Q1) prediction model. The prediction results of grey GM (1 ~ 1)-Markov model are compared. The results show that the success rate of the model is more than 66, which is superior to the grey GM (1t1) prediction model and the grey GM (1K1) -Markov prediction model, which shows that the prediction model established in this paper has good recognition accuracy and reliability.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.265

【参考文献】