一种基于流计算的铁路车流推算架构模型研究  

A Train Flow Estimation Architecture Model based on Stream Computing

作　　者：王斌[1] 金福才 谢玉霞 WANG Bin;JIN Fu-cai;XIE Yu-xia(Transport and Logistics School, Southwest Jiaotong University, Chengdu 610031, Sichuan, China;Institute of Computing Technologies;China Academy of Railway Sciences Corporation Limited, Beijing 100081, Chin)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都610031 [2]中国铁道科学研究院集团有限公司电子计算技术研究所,北京100081

出　　处：《铁道运输与经济》2018年第7期86-89,共4页Railway Transport and Economy

基　　金：中国铁路总公司科技研究开发计划课题(2017X009-C)

摘　　要：车流推算是铁路编制每日调度工作计划、进行日常车流组织的基础性工作。我国铁路车流输送距离长,运输组织复杂,车流状态变化频繁。通过分析车流推算过程及计算应用的处理特点,提出一种基于流计算的车流推算架构模型,将推算过程分解为多个子任务流进行并行处理。通过采用微应用计算架构,充分利用流计算在计算密集性、多级并行性、局域性的优势,可以实现实时车流推算和车流分析,以保证车流推算处理结果的及时性和有效性。Train flow estimation is the basic calculation job for making daily railway dispatching plans and organizing vehicle flows. The train travel distance is long in China, while the process of transport organization is complex, and the change of vehicle flow is frequent. The train flow estimation process and its calculation steps are analyzed, and a train flow estimation architecture model based on stream computing is proposed by dividing the estimation process into many small sub-task streams, which can be parallel processed. Using this micro-application architecture can take fully advantages of the stream computing in density, parallel processing and localization. The model can realize the real-time train flow estimation and analysis, and make the estimation results more timely and efficient.

关 键 词：铁路 车流推算 流计算 微应用架构 并行处理 架构模型 

分 类 号：U292.6[交通运输工程—交通运输规划与管理]