Abstract: In order to accurately describe the spatial-temporal evolution of the traffic flow and improve the accuracy of short-term traffic flow prediction, this paper builds the prediction model based on GM（1，N）-Markov chain, by fusing the spatial-temporal traffic information. Firstly, the prediction section and the association sections are taken as a grey system, and the grey relational analysis of the sections and setting the lowest threshold of grey correlation degree can mine the spatial information and reduce the ineffective information; secondly, the prediction section and strong correlation sections are analyzed systematically and then predicted by GM（1，N） model. In view of the failure caused by the randomness of sequence, the Markov chain is introduced to modify the forecasting model; finally, VISSIM is used for simulating, the simulation respectively taken 2 min, 5 min, 10 min as a time interval, the predicted average relative error is 9.30%, 5.95% and 3.20% respectively, the model accuracy is optimal, and this model is proven to be effective.

Key words: intelligent transportation, traffic flow forecasting, grey system, Markov chain

摘要： 为了准确描述交通流的时空演化过程并提高交通流短时预测的精度，融合时空交通流信息，即时间维度的交通流量信息和空间维度的路网耦合信息，构造基于GM（1，N）-Markov 链的组合预测模型。将预测路段与关联路段看作是一个灰色系统并对其进行灰关联分析，通过对灰关联度最低阈值的设定，实现了空间信息的深度挖掘和对无效信息的过滤清洗；利用多维GM（1，N）模型对预测点与强关联点作全局、系统的分析预测，并针对GM（1，N）对随机性较大的数列可能出现预测失效的问题，引入马尔科夫链对模型进行修正；利用VISSIM对模型进行仿真验证，分别以2 min、5 min、10 min为时间间隔进行仿真模拟，预测平均相对误差分别为9.30%、5.95%、3.20%，模型精度均为优，证实模型是有效的。

关键词: 智能交通, 交通流预测, 灰色系统, 马尔科夫链