高速公路施工控制区动态交通流预测的LSTM-BiGRU-Attention模型

LSTM-BiGRU-Attention model for dynamic traffic flow prediction in highway construction control zones

为提前准确预知高速公路施工控制区交通流变化趋势,解决交通流时间序列中的长期依赖问题,文中建立了高速公路施工控制区动态交通流预测的LSTM-BiGRU-Attention模型。首先,将预处理后的动态交通流数据集按时间步长顺序输入到LSTM网络,对交通流信息建模和学习。然后,引入BiGRU和Attention机制以更好地捕捉上下文信息和提供更具针对性的权重分配。最后,将构建的LSTM-BiGRU-Attention模型与其他模型进行交通流预测对比,评估模型性能。实验以G35济广高速公路某施工控制区交通运行情况为案例进行研究,结果显示该模型的平均绝对误差MAE为1.91,均方根误差RMSE为2.83,决定系数R^(2)为0.79,平均绝对百分数误差MAPE为3.23。对比其他模型,LSTM-BiGRU-Attention模型的4个评估指标均有所下降,说明该模型可为高速公路施工控制区提供更加精准的预测。

To achieve early and accurate prediction of traffic flow changes in construction control zones of highways and address the issue of long-term dependencies in traffic flow time series,we are establishing the LSTM-BiGRU-Attention model for dynamic traffic flow prediction in construction control zones.The model first processed the dynamic traffic flow dataset and sequentially fed it into the LSTM network,where it modeled and learned the traffic flow information.Next,the BiGRU and Attention mechanisms were introduced to better capture contextual information and provide more targeted weight allocation.Finally,the LSTM-BiGRU-Attention model is compared with other models for traffic flow prediction and evaluation of model performance.The experiments were conducted using the traffic operation situation in a construction control zone of the G35 Zhengzhou-Guangzhou Expressway as a case study.The results show that the model has an average absolute error MAE of 1.91,a root mean square error RMSE of 2.83,a coefficient of determination R 2 of 0.79,and an average absolute percentage error MAPE of 3.23.Compared to other models,the LSTM-BiGRU-Attention model shows a decrease in all four evaluation metrics,indicating that it can provide more accurate prediction models for construction control zones of highways.