基于二进制序列索引的灰色马尔科夫交通流量预测模型

Grey-Markov model for traffic flow forecast based on binary sequence index

针对短时交通流量预测在精度和收敛速度方面的不足,将二进制序列索引和灰色马尔科夫波动性预测模型相结合,用于短时交通流量预测.通过二进制序列索引,将传统灰色马尔科夫模型中的直接流量数据变为间接流量数据,以减少后期数据处理中误差导致的蝴蝶效应;同时通过波动性数据处理方法的逆方法将预测值还原到索引序列,使灰色马尔科夫预测模型可用于波动性数据的预测,并自主调整预测精度和收敛速度.以松原市兴业大街松原大路交叉口某一进口道的短时交通流量过程为例,对模型进行检验,结果表明:将经过二进制序列索引后的间接交通流量数据运用于优化后的灰色马尔科夫模型后,短时交通流量预测值的相对误差由0.46缩小至0.07,大幅提升了预测数据的精度,说明模型具有良好的预测精度,可以满足短时交通流量预测的要求,具有较高的实用性.

For the shortcomings of the short-term traffic flow prediction on precision and convergence speed,a method which combines the binary sequence volatility index and Grey-Markov prediction model was proposed to predict of the short-term traffic flow. The direct traffic data of traditional Grey-Markov model was changed into indirect traffic data by binary sequence index in order to reduce the later data processing error caused by the butterfly effect. Meanwhile the projec-tions were restored into the index sequence using the inverse method of processing method the data with volatility,and the prediction accuracy and convergence speed were actively adjusted. The model was tested with the short-term traffic flow of the intersection of Xinye street and Songyuan street in Songyuan city for example. The results show that relative error of the short-term traffic flow prediction is shortened from 0.46 to 0.07 by using the indirect traffic flow data which indexed by binary sequence for the optimized Grey-Markov model,and the prediction accuracy are significantly improved. The results suggest that the model has better prediction precision,and can satisfy the requirement of short-term traffic flow prediction with higher practicability.