高速铁路调度指挥计划同步优化模型和算法

While expectations from travelers on the punctuality of train services keep increasing, the traditional sequential approach for optimizing railway traffic management plans has shown its disadvantages, due to its negative impacts on the quality of plans. In view of this, this project aims to devote efforts to study models and algorithms for simultaneously optimizing railway traffic management plans.. This project will first investigate the characteristics of perturbations and passengers’ choice behaviors under delays, followed by building up formalized description models for passengers’ travel demand and resource capacities based on time-space networks. Then this project will put forward a simultaneous integer programming mathematical model for optimizing railway traffic management plans.. In order to speed up solution process and improve solution quality, this project will propose a rolling-horizon-based closed-loop optimization framework, in which Lagrangian relaxation and conflict-resolving based algorithms and a parallel computing framework will be designed and implemented. . Finally, a proto-type computer system will be designed and developed. The data for case studies will be prepared by collecting historical train monitoring data, tickets sales and questionnaire surveys. A series of comprehensive experiments will be conducted to test the proposed models and algorithms.. The implementation of this project will be beneficial for improving the quality of railway traffic management operations and providing theoretical guidelines and scientific basis for developing next generation advanced comprehensive railway traffic management computer systems on high-speed railway networks and meeting passengers’ expectations on safe, efficient and punctual train services.

随着人们对高速铁路出行准时性要求的不断提高，既有调度指挥计划“分步优化”理论呈现出一定的局限性，很大程度上限制了调度指挥的效果。本项目拟选择“高速铁路调度指挥计划‘同步优化’”这一国际前瞻性课题开展研究。首先，研究我国高速铁路初始扰动特征和非正常条件下旅客选择行为，基于时空网络建立旅客出行需求和行车资源的形式化描述模型，建立调度指挥计划同步优化编制整数规划模型；然后，为提高模型求解速度和质量，构建调度指挥计划滚动闭环优化框架，设计基于拉格朗日松弛和冲突疏解的求解算法，提出基于并行计算的求解机制；最后，研发计算机原型系统，使用统计和调查相结合的方式获取实例数据对项目提出的模型和算法进行验证。本研究将有助于提升调度指挥计划的质量，为研发新一代先进高速铁路综合调度指挥计算机系统，满足人们安全、高效、准时的出行需求提供重要的理论指导和科学依据。

随着人们对高速铁路出行准时性要求的不断提高，既有调度指挥计划“分步优化”理论呈现出一定的局限性，很大程度上限制了调度指挥的效果。本项目选择“高速铁路调度指挥计划‘同步优化’”这一国际前瞻性课题开展研究。首先，考虑动态的旅客出行需求，基于时空网络建立旅客出行需求和行车资源的形式化描述模型，建立调度指挥计划同步优化编制整数规划模型；然后，为提高模型求解速度和质量，设计基于拉格朗日松弛和冲突疏解的求解算法；最后，使用统计和调查相结合的方式获取实例数据对项目提出的模型和算法进行验证。本研究将有助于提升调度指挥计划的质量，为研发新一代先进高速铁路综合调度指挥计算机系统，满足人们安全、高效、准时的出行需求提供重要的理论指导和科学依据。