预测型列车运行调整计划鲁棒优化方法研究

Train dispatching (train traffic management) currently involves application of rule-based local dispatching measures for many lines and stations in case of minor/severe disturbances. Working in a predictive manner is poorly supported and train traffic controllers are usually restricted to just solving problems reactively after they occur (i.e. in a reactive manner). This often results in unnecessarily long delays and a decreased reliability of the railway transport system. With the increasing expectation of reliable train services and heavily utilized railway network worldwide, the current "Reactive" train dispatching methodology is insufficient to manage train traffic effectively. Minor disturbances (e.g. signal failure) may easily disorder the timetable and spread large delays all over the network..The goal of this project is to develop a new control approach, i.e., "Predictive and robust" methodology for managing networked train traffic more efficiently. We will first propose a conceptual model which describes the relations among sub-models. Based on analyzing the impacting mechanism of disturbances on normal train operations, we will put forward a model which can be used to predict potential train conflicts. Furthermore, we will present a scenario-based rolling horizon linear mixed integer programming mathematical model, based on a dynamic network flow formalization of train traffic management. Considering the great complexity of the proposed optimization model, we will design and implement a Lagrangian relaxation based parallel solution algorithm. Finally, we will develop a prototype computer system of predictive and robust train traffic management for testing the effectiveness of the proposed solution approach..The main theoretical contribution of this research is to speed up the research in train dispatching field. Moreover, considering the similarity between predictive train dispatching and dynamic job shop scheduling in a dynamic and stochastic environment with constrained resources, this research will also enrich and stimulate the research in the latter field..The main applicable value of this research is that predictive and robust train traffic management on the network level is a cost-efficient way to accommodate the expected growth of railway transport. It improves train traffic punctuality and travel time reliability, increases utilization of existing and future railway infrastructure, passenger satisfaction and social welfare, and reduces energy consumption and costs due to congestion. This research also provides a significant contribution for developing the new generation advanced train traffic control system.

随着人们对铁路出行准时性要求的不断提高，既有"反应型"列车运行调整的理论呈现出一定的局限性，很大程度上限制了列车运行调整的效果。本项目选择"'预测型'列车运行调整计划鲁棒优化"这一国际前瞻性研究课题，以建立该问题的理论框架模型为切入点，接着揭示外界扰动因素对正常列车运行秩序的作用机理，提出列车间"潜在"冲突的预测模型，在此基础上，建立面向多冲突情景的列车运行调整计划鲁棒优化数学模型，然后设计高效的并行求解算法，最后进行实例验证。本研究将有助于提升列车运行调整计划的预见性能、优化性能、适应性能和应用效果，推动铁路列车运行调整领域的理论研究，丰富动态工件调度研究领域的理论体系，同时，可为我国研发新一代先进列车调度指挥系统，满足人们安全、高效、准时的出行需求提供重要的理论指导。

随着人们对铁路出行准时性要求的不断提高，既有的“反应型”列车运行调整理论呈现出一定的局限性，很大程度上限制了列车运行调整的效果。本项目选择“‘预测型’列车运行调整计划鲁棒优化”这一前瞻性研究课题，首先建立了路网条件下列车运行调整计划同步优化模型，实现了列车到发时刻和列车运行路径的同步优化。随后，通过揭示外界扰动因素对正常列车运行秩序的作用机理，提出了列车间不确定性冲突的预测模型。基于列车运行调整计划同步优化模型和不确定性冲突预测模型，建立了面向多冲突情景的列车运行调整计划鲁棒优化数学模型。为了提高求解效率，模型采用基于拉格朗日松弛的算法进行求解，并基于该启发式算法研发了“预测型”列车运行调整计算机原型系统。最后，通过算例验证了模型的鲁棒性优化效果，实验结果表明，与CPLEX优化软件相比，所提出的基于拉格朗日松弛的求解算法具有高效性，能够快速收敛得到可行解，且解的质量可以衡量。本研究将有助于提升列车运行调整计划的预见性能、优化性能、适应性能和应用效果，推动铁路列车运行调整领域的理论研究，丰富动态工件调度研究领域的理论体系，同时，可为我国研发新一代先进列车调度指挥系统，满足人们安全、高效、准时的出行需求提供重要的理论指导。