基于多学科协同优化的铁路技术站作业计划整体编制研究

Technical stations are important terminals in the railway freight transportation network. Their operational efficiency significantly impacts the efficiency and quality of the total railway freight transportation service. The daily activities of a technical station are organized by designing an operation plan, in which several both independent and interrelated complicated decisions have to be optimized simultaneously. The operation plan of technical stations is still manually designed by experienced scheduler till now. Meanwhile, previous works cannot adequately satisfy the practical operational requirements. Hence, it is urgent to develop theoretically rigorous and practically implementable optimization theories and methods to holistically design the operation plan of technical stations. Based on the multidisciplinary collaborative optimization theory, this project explores the inherent decomposable characteristics of the operation plan and decomposes it under different operational conditions consistently into three critical sub-problems including the railcar allocation problem, the route scheduling problem and the shunting scheduling problem. Firstly, we employ the two-stage flow shop scheduling theory to develop optimization approaches for the railcar allocation problem with strict loading conditions. Secondly, we utilize the job shop scheduling theory to develop optimization approaches for the route scheduling problem with time-space consistency constraints. Thirdly, we refer the fixed interval scheduling theory to develop optimization approaches for the shunting scheduling problem with assembly sequence requirements. Finally, we combine the solution approaches of the three critical sub-problems and utilize the multidisciplinary collaborative optimization theory to develop global optimization approaches for the holistic operation plan design problem. The output of the project can provide certain theoretical guidelines and technological supports for achieving the automated, intelligent and efficient management of the operations in technical stations.

技术站是铁路货物运输网络中的重要节点，其工作组织水平显著影响整个铁路货物运输的服务效率和质量。技术站的日常工作通过编制作业计划来组织，需同时解决多项既相对独立又相互联系的复杂决策。作业计划至今仍由车站调度员凭借经验编制，且既有研究不能完全满足实际运营的要求，亟待开发理论缜密且实际可用的作业计划整体优化编制理论与方法。本项目基于多学科协同优化理论，利用作业计划本质上的可分解性，将不同运营条件下的作业计划整体编制问题统一分解为配流、进路调度和调车调度子问题。运用两阶段流水车间调度理论开发考虑严格满轴条件的配流优化方法；利用作业车间调度理论开发具有时空一致性约束的进路调度优化方法；借鉴固定区间调度理论开发囊括编组顺序要求的调车调度优化方法；联合子问题优化方法，开发基于多学科协同优化的作业计划整体编制全局优化方法。研究结果可为实现技术站工作组织的自动化、智能化和高效化提供一定的理论依据和技术支持。

技术站是铁路货物运输网络中的重要节点，其工作组织水平显著影响整个铁路货物运输的服务效率和质量。技术站的日常工作通过编制作业计划来组织，需同时解决多项既相对独立又相互联系的复杂决策。本项目基于多学科协同优化理论，利用作业计划本质上的可分解性，将不同运营条件下的技术站作业计划整体编制问题统一分解为配流、进路调度和调车调度子问题。.首先研究了单向单解单编运营条件（1个改编系统、1台解体调机、1台编组调机）下的配流问题，构建了1个混合整数线性规划模型，分别设计了1个拉格朗日松弛算法、1个启发式算法和1个基于整数编码的遗传算法；将所提出方法拓展于解决了单向并解并编、双向并解并编运营条件下的更复杂配流问题；对于各类运营条件下的实际规模的例子，所提出方法均能在3 min内获得最优解或者近似最优解。.其次研究了不同类型到发车场在不同制式联锁系统下的进路调度问题，将该问题巧妙转换为1个一般性的约束指派问题，构建了0-1线性规划模型，讨论了模型的可拓展性，并设计了迭代算法以解决模型可能的不可行问题；对于实际规模的例子，所提出的通用进路调度算法可在1 min内获得最优解。.接着研究了具有站顺要求的简单编组调车问题（1列车列编顺为1列车列），设计了1类位串法，开发了1个两阶段迭代搜索算法；将所提出的位串法拓展来解决了更为复杂的将多列车列编顺为多列车列的批量编组调车问题；对于实际规模的简单和批量编组调车例子，位串法能在1 min内收敛。研究了考虑解编调车协调的调车线运用问题，构建了1个0-1线性规划模型，设计了1个两阶段分解算法，该算法能在1 min内将实际规模例子求解到最优。.最后，基于已提出的配流、进路调度和调车调度优化建模策略，构建了基于多学科协同优化的技术站作业计划整体编制全局优化框架。.项目研究成果可为提高技术站作业计划质量、实现技术站作业计划编制的自动化、智能化和高效化提供理论依据和决策支持。