超大型集装箱码头不同于中小型码头，具有平面运输路网复杂、设备众多且类型多样等典型特性，这些因素会对整体调度效率产生显著影响。复杂的平面运输路网要求更精细化的全场路径规划，否则可能导致集卡运输全局效率低下。多样化和庞大的设备集群，需要协调不同设备的调度需求和控制方式，以确保能够协同管控各类设备高效运行，实现安全和效率的平衡。基于超大型集装箱码头的典型特征分析，本论文依托宁波舟山港梅山港区作为应用场地，详细研究了集装箱码头调度系统中存在的关键问题，提出了智能调度系统的理论创新体系架构和关键技术，并进行了工程化创新应用和效果验证。为适应超大型集装箱码头的调度需求，建立了一套双维智能调度体系架构，横向集成采用“云-网-边-端”一体化设计，纵向集成则实现了从数智引擎到业务应用的端到端闭环，有效支撑了集装箱码头的数智化调度水平。针对内、外集卡的定位问题分别设计了定制化解决方案。内集卡通过基于多模型无迹卡尔曼滤波器的融合定位算法，能够实时整合三网定位数据，提高了定位稳定性。外集卡的定位则基于车路协同系统展开，在港区构造“智联的路”，从而提供感知定位、位置识别和事件识别等能力，支撑了调度系统对定位系统稳定性和精准性的需求。针对局部调度不足，提出一个全局任务调度算法框架及模型。算法框架以双层算法为核心，实现了最优路径规划和实时协同管控。模型则协同算法框架，以时间成本最优为目标，生成集装箱最优作业序列。 为实现对码头各类资源的精准配置，构建了具有“四层服务”的数字孪生框架展示集装箱码头运营情况，并通过专题视图进行多角度监控和分析。此外，创新性地引入集卡仿真还原技术，帮助调度员挖掘历史作业瓶颈信息，优化调度策略。本论文针对宁波舟山港梅山港区的智能调度系统进行了实施和验证，取得了显著的成果。一方面，基于数字孪生平台对包含感知专题在内的五大专题进行实船应用效果的可视化验证；另一方面，本论文对码头效能评估的六大指标进行数据分析验证。具体为： 岸桥单机效率提升7.3%，集卡效率提升4.0%，场桥小时作业量提升6.6%。此外，集卡作业等待时间降低57%，远控场桥指令拒绝率降低了34.6%，集卡重载率提升1.8%。这证实了所提出的超大型集装箱码头智能调度系统在提升作业效率方面的显著贡献。

Mega Container Terminals differ from medium and small terminals in that they have complex surface transportation networks, numerous and diverse equipment, among other typical characteristics. These factors significantly affect the overall scheduling efficiency. The complexity of the surface transportation network demands more detailed and precise path planning across the entire terminal, otherwise, it could lead to low global efficiency in truck transport. The diverse and large equipment clusters require coordination of the scheduling demands and control methods of different equipment to ensure the efficient operation of all types of equipment, balancing safety and efficiency. Based on the typical characteristics of Mega Container Terminals, this paper takes the Meishan Port area of Ningbo Zhoushan Port as an application site and thoroughly investigates the key issues in container terminal scheduling systems. It proposes a theoretical innovation system architecture for the intelligent scheduling system, key technologies, and engineering applications, followed by effect verification.To meet the scheduling demands of Mega Container Terminals, a dual-dimensional intelligent scheduling system architecture has been established. The horizontal integration adopts a “cloud-network-edge-end” integrated design, while vertical integration realizes an end-to-end closed loop from digital intelligence engine to business application, effectively supporting the digital and intelligent scheduling level of container terminals.For the positioning issues of internal and external trucks, customized solutions were designed. Internal trucks use a fusion positioning algorithm based on a multi-model unscented Kalman filter, which integrates three-network positioning data in real-time to improve positioning stability. External truck positioning is based on the vehicle-road collaborative system, where a "smart connected road" is constructed in the port area, providing capabilities for perception positioning, location recognition, and event recognition, thus supporting the scheduling system's need for stability and accuracy in positioning systems.A global task scheduling algorithm framework and model are proposed to address local scheduling inadequacies. The algorithm framework includes customized solutions for container selection, vehicle selection, container handling, and gantry crane loading/unloading, with an emphasis on the coordination between operation processes. The model comprehensively considers both cost and efficiency during the scheduling process, achieving a balance between resource scheduling and cost control.To achieve precise allocation of terminal resources, a digital twin framework with a “four-layer service” is built to display the operation status of the container terminal and enable multi-angle monitoring and analysis through specialized views. Furthermore, an innovative container truck simulation technology is introduced to help schedulers uncover historical operational bottlenecks and optimize scheduling strategies.This paper implements and verifies the intelligent scheduling system at the Meishan Port area of Ningbo Zhoushan Port, achieving significant results. On one hand, the digital twin platform provides a visual verification of the real-ship application effects across five key areas, including perception. On the other hand, the paper analyzes six performance evaluation indicators of the terminal through data verification. Specifically: bridge crane efficiency improved by 7.3%, truck efficiency improved by 4.0%, and gantry crane hourly workload increased by 6.6%. Additionally, truck operation waiting time decreased by 57%, remote-controlled gantry crane command rejection rate dropped by 34.6%, and truck heavy load rate increased by 1.8%. These results confirm the significant contribution of the proposed intelligent scheduling system for Mega Container Terminals in improving operational efficiency.