数据驱动的港口群水路集疏运服务优化策略研究

水路运输具有运量大、成本低、能耗少、可靠性高等优势，在综合运输系统中扮演着重要的角色。本项目依托港口物流多源数据，以水路集疏运服务系统为研究对象，针对服务定价策略和调度机制，建立双层规划模型探究利益主体之间的交互关系，并利用数据驱动的方法，研究动态定价与调度策略，进一步降低水运服务的运营成本、提高港口企业的效益、增加服务的弹性。首先，基于多源数据挖掘水运需求的动态性和运输系统的特点，并利用机器学习方法对集疏运需求进行预测；其次，基于港口企业与货主之间的交互关系构建双层规划模型，分析最优定价和调度决策，并考虑港口群企业之间的竞合关系，建立双重博弈模型；然后，针对水路集疏运系统的不确定性，构建基于滚动计划的动态定价和调度模式，并结合运输需求、港口状态、船舶状态等实时信息，研究实时定价和调度决策以及设计在线优化算法；最后，将研究成果在具有典型水路集疏运系统的枢纽港企业进行应用和效果验证。

Waterway transport plays an essential role in freight transport industry through providing high capacity and reliable transport service, reducing carbon emission and cost, and relieving traffic congestion. This project studies the dynamic transport service pricing and scheduling problem in a waterway transport system based on multi-source data in port logistics, with the objective of improving the profit of port enterprises, reducing the transportation cost, and improving transport service resilience. Specifically，this project will first conduct data mining regarding the dynamics of the transport demands as well as the characteristics of the waterway transport system, and make demand forecast by using machine learning methodologies. Second, a bilevel model with one port enterprise as the leader and multiple shippers as followers will be developed for obtaining the optimal transport service pricing and scheduling, and a dual game theoretic model will be formulated considering the co-opetition in port clusters. Third, regarding the uncertainty in waterway transport systems, a dynamic pricing and scheduling scheme will be proposed based on the rolling horizon approach; and the online pricing and scheduling decisions and optimization algorithms will be designed based on the real-time data including transport demand changes and the conditions of ports and vessels. Finally, the analytical and experimental results will be extensively tested based on the data provided by the main port enterprises with typical waterway transport networks.