摘要
铁路选线是一个复杂的工程问题,尤其是当稠密约束选线环境内存在起伏地形高差约束和大量障碍物时,现有计算机辅助线路设计方法往往需要耗费大量的计算时间和资源才能生成优化方案,甚至容易停滞,无法找到可行方案。为解决这一问题,提出一种三维快速搜索随机树算法以期快速生成满足所有约束的线路方案。首先,为避免随机树搜索陷入局部最优,提出平-纵整合式随机树启发式采样方法,将随机树搜索扩展到三维空间并实现对选线搜索区域的全面探索。其次,为了在线路搜索过程中高效提取相关环境信息,提出多源异构综合地理信息的统一管理方法,针对环境信息特点定制相应的储存策略,将地形、障碍物等环境信息离散到综合地理信息模型中,并在随机树搜索过程中动态处治障碍物约束。随后集成启发式采样方法和约束处治算子,提出随机树进化搜索方法,在随机树拓展过程中高效检索和处治障碍物,快速生成优化路径方案。最后,将此方法应用于一个真实稠密约束线路案例中,实验结果证明,此方法能实现对所有障碍物约束的空间绕避,并能快速产生优化线路方案,相比于人工方案,机选方案的造价降低了4.8%。实验结果表明此方法可以提高线路设计效率,为人工设计提供有价值的参考。
Railway design presents a complex engineering challenge,particularly in areas with undulating terrains and numerous obstacles,in the study areas.Existing computer-aided alignment optimization methods often require extensive computation time and resources to generate an optimized alignment and are even prone to stagnate and fail to find a feasible alignment.To address these issues,a 3D rapidly exploring random tree algorithm was proposed for rapidly developing feasible alignments that satisfy all the constraints.First,a random tree heuristic sampling method termed as the horizontal-vertical integrated sampling method was proposed for avoiding the random tree search process falling into a local optimal.This sampling method extends the random tree search to three-dimensional spaces and achieves a comprehensive exploration of the entire study area.Second,to efficiently retrieve relevant environmental information during the alignment search process,a unified management method for multi-source heterogeneous integrated geographic information was proposed.Specifically,based on the characteristics of different environmental information,such as terrains and obstacles,specific storage strategies were designed for discretizing them into a comprehensive geographic information model.Then,this environmental information was capable to be dynamically checked during alignment search process.Afterward,the random tree evolutionary search method was obtained by integrating the heuristic sampling method and the obstacle handling operator.This method can efficiently retrieve and handle relevant obstacles during search process as well as to rapidly generate optimized path solution.Ultimately,the proposed method was applied to a real-world case in a study area with multiple constraints.The experimental results reveal that this method can effectively bypass all the obstacle constraints and rapidly generate an optimized alignment.As compared to the best manually designed alignment solution,the optimized alignment derived from the proposed method achieves 4.8%lower construction cost.The experimental results show that this method improves alignment design efficiency and provides valuable references for human designers.
作者
万昕洁
蒲浩
冉杨
李伟
胡建平
乔俊飞
WAN Xinjie;PU Hao;RAN Yang;LI Wei;HU Jianping;QIAO Junfei(National Engineering Research Center of High-speed Railway Construction Technology,Changsha 410075,China;China Railway Group Limited,Beijing 100039,China;School of Civil Engineering,Central South University,Changsha 410075,China;China Railway Eryuan Engineering Group Co.,Ltd.,Chengdu 610031,China;China Railway Engineering Design Consulting Group Co.,Ltd.,Beijing 100073,China)
出处
《铁道科学与工程学报》
EI
CAS
CSCD
北大核心
2024年第8期3142-3152,共11页
Journal of Railway Science and Engineering
基金
中国中铁股份有限公司科技研究开发计划项目(2022-重大-20)
国家自然科学基金资助项目(52078497)
国家“十四五”重点研发计划项目(2021YFB2600400)。
关键词
铁路设计
线路优化
启发式采样
快速搜索随机树
约束优化
railway design
alignment optimization
heuristic sampling process
rapidly exploring random tree algorithm
constrained optimization
