论文部分内容阅读
基于城市道路交通流按日周期变化的数据特征,提出一种可自动选择步长的灰色模型算法.将其应用到美国Minnesota的两个道路交通流的预测,并和传统灰色模型、历史平均法以及滑动平均法对比.数值实验结果表明:改进的灰色模型能够大幅降低预测绝对误差,预测精度高,稳定性好,适用于城市道路短时交通流的实时预测.
Based on the data characteristics of urban road traffic flow varying according to the daily cycle, a gray model algorithm with automatic step size selection is proposed, which is applied to the prediction of two road traffic flows in Minnesota, United States, and compared with the traditional gray model, historical average method And the sliding average method are compared.The numerical results show that the improved gray model can significantly reduce the absolute error of prediction, and has high prediction accuracy and good stability, and is suitable for real-time prediction of urban road short-term traffic flow.