高速铁路多列车票额分配计划协同优化研究

With the development of high-speed railway network, the service level had been improved remarkably as the benefit of network operation, which makes the demand increase strongly. Seat allocation with high quality could provide significant guarantee for improving the service of passengers and enhancing operation efficiency of high-speed railway system. As the train flows will vary with the changes of line planning, schedule and seat allocation scheme, it is difficult to predict the train flows. The traditional optimization method of seat allocation in the existing railway system cannot solve the problem of high-speed railway considering the time-space passenger demands, with the neglect of ticket-booking competing behavior and the Master-slave game relationship between railway operator and travel passengers. This project designs the method of estimating the time-varying demand of high-speed railway and analyze the characteristic of ticket-booking. Based on the behavior of ticket-booking, we analyze the choices of booking time and tickets, builds the time-space network of multi-train cooperation, discuss the competition equilibrium of ticket-booking, and design the passenger assignment method of high-speed railway system considering the time-varying demand and the strict capacity constraints of trains or tickets. This project also using the bi-level programming to describe the Master-slave game relationship between railway operator and travel passengers, and optimize collaboratively the multi-train seat allocation from the interests of passengers, railway company and both respectively, in order to obtain more reasonable and acceptable scheme of seat allocation for high-speed railway system. This project strives to establish the method theory of coordinating optimizing of multi-train seat allocation plan for high-speed railway network.

随着高速铁路快速发展，网络化运营大幅提升服务水平，同时带来高速增长的需求。票额分配是提升高速铁路旅客服务和运营效率的重要保障。然而，列车客流随开行方案、运行图和票额分配计划的变动而变化，利用历史数据难以预测；同时忽视旅客购票竞争行为以及与铁路企业之间的主从博弈关系，会导致传统票额分配方法难以适用于高速铁路时空客运需求。本项目利用高速铁路售票数据，分析时变需求的获取方法和购票特征。针对高速铁路旅客的购票行为，考虑时变需求和严格列车能力（或票额限制）约束，分析旅客的购票时间和购票选择决策，构建多列车时空协同网络，探讨旅客购票竞争均衡，建立高速铁路网络均衡分配方法。采用双层规划模型，描述旅客与铁路企业之间的主从博弈关系，分别从旅客、铁路企业、以及兼顾双方利益角度，进行多列车票额分配协同优化，探讨更合理和更易接受的票额分配优化方案。本项目力图建立高速铁路多列车票额分配计划协同优化理论。

高速铁路网络运营具有高速度和高密度开行特征，使得高铁需求具有明显的时变性。本项目以高速铁路售票数据为基础，分析了旅客时变需求特征和旅客的提前购票选择行为，并探讨了高铁旅客与铁路企业的主从博弈关系，考虑双方利益综合优化票额分配方案，系统研究了高速铁路多列车票额分配计划协同优化理论和方法。.旅客出行需求是优化票额分配和制定售票计划的基本输入数据。本项目分析高速铁路旅客购票数据特征，研究了高速铁路时变需求预测方法，设计了基于客流量的高铁旅客时变需求分布逆分配估计法和基于购票记录的最大熵估计模型，以获得高铁需求时变分布；并提出了高速铁路日常客运量VMD-GA-BP短期预测方法和基于双层平行小波神经网络模型的中期客流预测方法，分别进行中短期日总客流量预测。.基于时变需求，本项目研究了高速铁路网络客流均衡分配方法。基于高铁列车时刻表构建了旅客出行时空网络,分析了旅客提前购票选择，构建了基于票额共用和票额分配方案的旅客提前购票竞争和路径选择均衡模型，并利用Danzig-Wolfe分解算法和时空网络无圈特性，设计了求解大规模网络能力的客流分配算法。同时考虑旅客出行可靠性，进一步提出了基于可靠性的高铁客流分配方法。.基于上述工作，从旅客和铁路企业两方面利益着想，在满足需求前提下，通过优化票额分配方案，最小化旅客出行总费用并降低旅客购票竞争费用，同时最大化铁路企业收益，构建了高速铁路多列车票额分配协同优化的双层规划模型；依据模型特征，将其转换为混合整数线性规划问题，设计了一种全局优化算法，最终提出一整套高速铁路多列车票额分配计划协同优化方法。最后进一步扩展，研究了基于收益管理的高速铁路票额票价协同优化方法，并分别设计求解算法。.项目各项工作在我国高速铁路实际案例进行了验算分析，效果良好。本项目建立了高速铁路多列车票额分配计划协同优化研究的系统理论，可以科学合理地优化高速铁路资源配置并提高运营效率，为高速铁路系统运营组织提供必要的理论支撑。