交通运输系统工程与信息 ›› 2023, Vol. 23 ›› Issue (5): 258-267.DOI: 10.16097/j.cnki.1009-6744.2023.05.027

• 运输组织优化理论与方法 • 上一篇    下一篇

考虑司机偏好的城市轨道交通混合乘务轮转模型

潘寒川1,戚博洋1,胡华*1,康磊2,沙悦1,刘志钢1   

  1. 1. 上海工程技术大学,城市轨道交通学院,上海 201600;2. 上海磁浮交通发展有限公司,上海 201204
  • 收稿日期:2023-06-02 修回日期:2023-07-18 接受日期:2023-08-02 出版日期:2023-10-25 发布日期:2023-10-23
  • 作者简介:潘寒川(1986- ),男,安徽青阳人,副教授,博士
  • 基金资助:
    国家自然科学基金(72371154,52072235);中国博士后科学基金(2021M692413)

Preference-oriented Task-type-mixed Crew Rostering Optimization Model for Urban Railway Transit

PAN Han-chuan1, QI Bo-yang1, HU Hua*1, KANG Lei2, SHA Yue1, LIU Zhi-gang1   

  1. 1. College of Urban Railway Transit, Shanghai University of Engineering and Science, Shanghai 201600, China; 2. Shanghai Maglev Transportation Development Co. Ltd, Shanghai 201204, China
  • Received:2023-06-02 Revised:2023-07-18 Accepted:2023-08-02 Online:2023-10-25 Published:2023-10-23
  • Supported by:
    National Natural Science Foundation of China (72371154,52072235);China Postdoctoral Science Foundation (2021M692413)

摘要: 基于非循环轮转方案,本文提出混合乘务轮转模式下的建模方法,并将司机对乘务计划的需求总结归纳为司机偏好,给出偏好的量化计算方法,加入乘务轮班计划求解目标,构建考虑司机偏好的城市轨道交通混合乘务轮转优化模型。设计基于大规模领域搜索的模拟退火算法,以上海地铁某线路的实际运营数据为背景进行案例分析。研究结果表明,计算结果与传统模型对比下,本模型所得的轮转计划实际出勤点偏好满足率约为90%,满足任务类型偏好分配率超过65%,优于传统轮转方案,且平均工作时长方差也有明显下降,计算结果验证了本模型的有效性和实用性。

关键词: 城市交通, 混合乘务轮转, 模拟退火算法, 司机偏好, 大领域搜索

Abstract: This paper proposes a modeling method for the mixed crew rotation mode based on an acyclic rotation scheme. It incorporates the driver's demand for the crew plan as their preference and provides a quantitative calculation method for this preference. Additionally, the crew rotation plan is included as a solution goal, creating a mixed crew rotation optimization model for urban rail transit that considers the driver's preference. The paper also designs a simulated annealing algorithm based on large-scale domain search to solve the model. The case study is conducted using actual operation data from a line in the Shanghai subway. The results demonstrate that the calculated results are improved when compare to those obtained from the traditional model. The optimized rotation plan achieves an actual attendance point preference satisfaction rate of approximately 90% , and the task type preference distribution rate exceeds 65%. These figures are considerably higher than those achieved by traditional rotation schemes. In addition, the model significantly reduces the average working time variance. The calculation results validated the effectiveness and practicality of the proposed model.

Key words: urban traffic, task-type-mixed crew rostering, simulated annealing algorithm, driver preference, large neighborhood search

中图分类号: