| 摘 要: 集装箱码头作为全球贸易物流网络的关键枢纽,其核心装卸设备——桥吊的作业效率,直接影响船舶在港停时及码头整体吞吐能力。通过对现有相关技术与文献的系统分析可发现,桥吊作业效率指标仅能在全部作业工序完工后获取,不具备实时反馈能力,难以满足调度环节的即时评估需求；现有码头作业预测研究大多以作业最终产出为预测对象,尚未建立面向作业运行全过程的动态效率预测方法,无法为一线操作人员的精细化、实时化调度提供有效支撑。针对上述问题,本文基于某集装箱码头真实数据,提出一种LSTM-Transformer融合模型的桥吊作业效率预测方法。首先,以船舶靠离泊时间为基准,将作业过程划分为连续时间切片,构建时间序列片段；进而,通过多步滞后相关性分析,挖掘各特征对桥吊效率的滞后影响规律,确定最优输入窗口,随后输入到LSTM-Transformer融合模型。实验结果表明,该方法使得预测精度相较于单一LSTM模型分别提升了21.3%和11.3%。 |
| 关键词: 时序特征 桥吊效率预测 滞后相关性分析 LSTM-Transformer |
|
中图分类号:
文献标识码:
|
| 基金项目: 浙江省自然科学基金青年基金(LQN25F030019) |
|
| Development of Time-Series Forecasting Model for Quay Crane Operational Efficiency at Container Terminals and Performance Analysis |
|
WANG Tengfei1, LIN Binxun2, FENG Xiaobo3, HU Kehan4, ZONG Yang5
|
1.School of Information Science and Engineering (School of Cyberspace Security), Zhejiang University of Science and Technology;2.School of Computer Science and Technology (School of Artificial Intelligence), Zhejiang University of Science and Technology;3.School of Information Science and Engineering, Ningbo Institute of Technology, Zhejiang University;4.Ningbo Zhoushan Port Nonferrous Ore Storage and Transportation Co., Ltd.;5.Technology and Information Engineering Department of Tianjin Port Co., Ltd.
|
| Abstract: As a critical hub in the global trade logistics network, the operational efficiency of quay cranes—the core loading and unloading equipment in container terminals—directly affects vessel turnaround time and the overall throughput capacity of the terminal. Through a systematic analysis of existing relevant technologies and literature, it can be observed that quay crane efficiency indicators are only available after all operation processes are completed, lacking real-time feedback capability and thus failing to meet the immediate evaluation requirements of scheduling. Most existing studies on terminal operation prediction focus on the final operational output and have not yet established a dynamic efficiency prediction method for the entire operation process, which makes it difficult to provide effective support for refined and real-time scheduling of frontline operators. To address the above issues, this paper proposes a quay crane operational efficiency prediction method based on a hybrid LSTM-Transformer model, using real-world data from a container terminal. First, using vessel berthing and departure times as references, the operation process is divided into continuous time slices to construct time-series segments. Then, multi-step lag correlation analysis is performed to uncover the lagged influence patterns of various features on quay crane efficiency and to determine the optimal input window, which is subsequently fed into the hybrid LSTM-Transformer model. Experimental results show that the proposed method improves prediction accuracy by 21.3% and 11.3% compared to a single LSTM model. |
| Keywords: temporal characteristics quay crane efficiency prediction lag correlation analysis LSTM-Transformer |
