线路-结构物-环境耦合约束下的铁路点线协同优化理论与方法

The railway alignment design standards are greatly promoted for the next-generate railway system namely high speed epoch. At the same time, the construction locations transform from plains in the east of China to mountainous areas in the west of China. The challenge arising in the situation is that how to cooperatively optimize high quality alignment and distribution of structures (stations, bridges and tunnels) concurrently in such a complexity environment. Unfortunately, the existing methods of railway alignment optimization cannot work out the cooperative optimization solution. The basic reason is that they ignore the inner relationships among alignment, structures and environment and their dynamic features. Our idea is to link 'point', i.e. station, bridge and tunnel, with 'alignment' as a whole. This linked action forms an innovative alignment-structure-environment coupled constraints (ASEcc) which establish a foundation in our railway alignment optimization. Considering the ASEcc is highly dimensional in nature, a hyper-graph semantic model is proposed to describe those constraints. Based on the ASEcc, the "point-alignment" cooperative intelligent search process is applied to achieve the optimization solution. This search process includes two basic steps: firstly, the feasible areas for route corridors are constructed based on hyper-graph semantic model in the basic geographic environment. According to composable criterion, feasible areas finally form non-inferior solution space of route corridors which constraints the final solution space into a group of feasible directions. Secondly, based on the route corridor, an adaptive distribution method for control points of alignment and a heuristic-based search method are proposed under ASEcc. Additional, the mechanism for dynamically creating and updating the coupling constraints is set up. Differencing from the traditional railway alignment optimization methods which considered as isolated, i.e. the "point" and "alignment" is disconnected, in static constraints, the contributions of this project are to rethink "point-alignment" coordination optimization under ASEcc which try to thoroughly solve the scientific problem that current railway alignment optimizations cannot be applied in the complex environment.

我国铁路由普速向高速发展，铁路线路设计标准大幅提升，建设区域由东部平原向西部山区转变，面临的环境日趋复杂，这些变化使铁路空间线位和结构物布设的协调控制难题变得尤为突出。现有智能选线方法忽视了线路，结构物，环境之间的耦合关系和动态特征，常常无法给出点线综合协调的优化解。本项目以"点线结合"为切入点，分析线路、结构物、环境之间的耦合约束关系，构建超图语义模型来刻画其多维语义关系；提出顾及线路-结构物-环境耦合约束关系的铁路点线协同智能搜索方法：应用超图语义模型，搜索线路走廊带可行区域，通过可组合性判别，形成线路走廊带的非劣解空间；在此基础上，研究耦合约束下控制点自适应分布方法，建立耦合约束的动态创建与更新机制，实现点线协调方案的启发式搜索。本课题突破传统静态约束下点或线的孤立、片面优化，重新考虑线路-结构物-环境耦合约束下的点线协同优化，力图从源头上解决铁路线路优化复杂环境受限的科学问题。

我国铁路由普速向高速发展，铁路线路设计标准大幅提升，建设区域由东部平原向西部山区转变，面临的环境日趋复杂，这些变化使铁路空间线位和结构物布设的协调控制难题变得尤为突出。现有智能选线方法忽视了线路，结构物，环境之间的耦合关系和动态特征，常常无法给出点线综合协调的优化解。本项目针对线路-结构物-环境耦合约束下铁路点线协同优化问题开展研究：建立了由数字地形、地质、地物、单价、结构物设计参数等信息组成的综合地理信息模型，为铁路点线协同优化奠定基础；提取线路、结构物、环境三者及三者间的空间语义关系，并抽象成数学表达式，构建数学模型准确刻画出线路-结构物-环境高维、动态、耦合的约束关系；提出了耦合约束的前-中-后三阶段处理方法，实现了耦合约束的动态处理；建立了综合考虑工程造价、运营费用、环境影响的铁路点线协同优化模型，并提出了双向距离变换与改进遗传算法相结合的搜索算法解算该模型，实现了复杂地形、地质环境下铁路线路、结构物的协同优化。基于本研究提出的理论和方法，开发了“铁路三维空间智能选线系统”，该系统已在川藏铁路、南美洲两洋铁路等复杂艰险山区得到成功应用。项目执行期间培养博士研究生6名，硕士研究生26名，其中已毕业博士研究生1名，硕士研究生15名；出版专著1部；发表科技论文16篇，其中SCI期刊检索4篇，EI期刊检索7篇，核心期刊检索4篇；申请国家发明专利4项，其中3项已授权；获软件著作权4项。本项目的成果可为铁路线路智能设计提供理论基础与技术支持。