首页|考虑供需优先特性的JSQ型空车调配优化模型及算法

考虑供需优先特性的JSQ型空车调配优化模型及算法

扫码查看
面向铁路商品汽车运输JSQ型车辆的空车调配问题,首先总结其调配计划制定-执行流程,将站点装车重要程度、空车库存及缓冲、路局存车限制等要素抽象为出空、需空优先特性,结合路网点线能力制约,提炼JSQ型空车调配问题关键要素;其次,构建供需网络及网络流模型,给出各优先特性的参数标定条件,实现JSQ型空车调配问题的数学描述;然后,通过拉格朗日松弛将原问题松弛为最小费用最大流问题,设计基于Ford Fulkson方法及优先队列优化的Bellman Ford算法求解松弛问题;最后,基于中铁特货跨局调配数据验证模型及算法有效性、基于中铁特货历史数据完成全路规模下的压力测试,通过与商用求解器的对比,表明本文方法能够准确、高效地实现优化.
Optimization Model and Algorithm for JSQ Empty Car Distribution with Priorities of Supplies and Demands
To reduce cost and increase efficiency in JSQ empty car distribution,our works are fourfold:Firstly,the de-signing-executing process of empty car distribution in China Railway Special Cargo Logistics Co.,Ltd(CRSCL)was summarized.The importance of loading stations,the inventory and the storage limit of empty cars were abstracted into supplying and demanding priorities.Combined with the section and the station capacity,the key factors in the empty car distribution problem were extracted.Secondly,a mathematical model and the priority parameter conditions were proposed based on the supply-demand network and network flow model,realizing the mathematical description of the problem.Thirdly,based on the use of a Lagrangian relaxation to relax the model into a minimum cost and maximal flow problem,the relaxed problem was solved by Ford Fulkson method and Bellman Ford algorithm with the priority queue.Finally,the effectiveness of the model and algorithm was verified according to the cross-bureau distribution data in CRSCL and the stress testing was done based on the historical data of CRSCL.The results show that the proposed method can tackle the problem precisely and efficiently,compared with the commercial solver.

empty car distributionJSQ railway carspriorities of supplies and demandsnetwork flow problem

张家瑞、李海鹰、王莹、冀柯维、罗涛

展开 >

北京交通大学轨道交通控制与安全国家重点实验室,北京 100044

北京交通大学交通运输学院,北京 100044

中铁特货物流股份有限公司市场部,北京 100055

空车调配 JSQ型车 供需优先特性 网络流问题

中铁特货物流股份有限公司科技研发计划

ZTTH-2022-WT-032

2024

铁道学报
中国铁道学会

铁道学报

CSTPCD北大核心
影响因子:0.9
ISSN:1001-8360
年,卷(期):2024.46(5)
  • 11