集装箱港区集卡到港量组合预测

Combined Forecast of Arrival Volume of Container Trucks in Container Port Area

摘要: 集装箱码头作业计划与每日的抵港集装箱卡车数量有关，为准确预测每日的集卡到港量，提高码头作业效率，本文将船期表和码头堆存费率等作为影响因素，建立集卡到港量短期预测的ARIMA-BP神经网络组合预测模型，并与不同模型的预测精度进行对比。以上海外高桥二号码头为例，验证了基于船舶的开截港时间与码头堆存费率等为影响因素的ARIMA-BP神经网络组合预测模型的有效性。结果表明，组合预测方法较单一模型具有更好的预测效果。在各种组合模型中，ARIMA-BP神经网络组合预测模型具有较高的精度和稳定性，是对集卡到港量进行短期预测的一种有效方法。

Abstract: The operation plan of a container truck is closely related to the number of arriving container trucks on a daily basis. In order to accurately predict the daily volume of trucks arriving at the port and improve the efficiency of terminal operations, this paper considers factors such as the shipping schedule and terminal storage rates. A combined prediction model of ARIMA-BP neural network is established for short-term forecast of container trucks at the port, and its forecast accuracy is compared with that of different models. Taking Shanghai Waigaoqiao Terminal II as an example, the effectiveness of the ARIMA-BP neural network combined prediction model, which takes factors such as vessel arrival and departure times and terminal storage rates into consideration, is validated. The results show that the combined forecasting method outperforms the individual models in terms of forecast accuracy. Among various combined models, the ARIMA-BP neural network model demonstrates higher precision and stability, making it an effective approach for the short-term forecast of container truck at the port.