• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2024, Vol. 60 ›› Issue (6): 58-68.doi: 10.3901/JME.2024.06.058

• 特邀专栏:数据-知识混合驱动的智能制造系统 • 上一篇    下一篇

扫码分享

大数据驱动的快消品终端拜访“云-边”联动决策与优化

赵阔1, 王皂琦1,2, 潘臻信1, 潘扬华3, 张中飞4, 屈挺1,3   

  1. 1. 暨南大学智能科学与工程学院 珠海 519070;
    2. 暨南大学信息科学技术学院 广州 510000;
    3. 暨南大学物联网与物流工程研究院 珠海 519070;
    4. 暨南大学管理学院 广州 510000
  • 收稿日期:2023-04-10 修回日期:2023-12-21 出版日期:2024-03-20 发布日期:2024-06-07
  • 通讯作者: 屈挺,男,1979年出生,博士,教授,博士研究生导师。主要研究方向为制造物联网、生产物流及供应链管理,工业工程。E-mail:quting@jnu.edu.cn
  • 作者简介:赵阔,男,1977年出生,博士研究生。主要研究方向为机械工程、物联网、大数据、人工智能、区块链。E-mail:zhaokuo@jnu.edu.cn;王皂琦,男,1998年出生,主要研究方向为人工智能。E-mail:wangzaoqi98@163.com;潘臻信,男,1998年出生,主要研究方向为人工智能。E-mail:pan_zhenxin_apply@163.com
  • 基金资助:
    2019年度“广东特支计划”本土创新创业团队(2019BT02S593)、国家自然科学基金(51875251)、 2018年度广州市创新领军团队(201909010006)和中央高校基本科研业务费专项资金(11618401)资助项目。

Big Data-driven FMCG Terminal Visits "Cloud-side" Based Decision-making and Optimization

ZHAO Kuo1, WANG Zhaoqi1,2, PAN Zhenxin1, PAN Yanghua3, ZHANG Zhongfei4, QU Ting1,3   

  1. 1. School of Intelligent Systems Science and Engineering, Jinan University, Zhuhai 519070;
    2. College of Information Science and Technology, Jinan University, Guangzhou 510000;
    3. Institute of Physical Internet, Jinan University, Zhuhai 519070;
    4. School of Management, Jinan University, Guangzhou 510000
  • Received:2023-04-10 Revised:2023-12-21 Online:2024-03-20 Published:2024-06-07

摘要: 随着我国的快消品消费市场快速增长,快消品公司的终端拜访成本显著增加,同时终端拜访场景面临着终端信息的实时获取、动态的决策与客户需求动态响应的问题。提出针对快消品终端拜访问题的“云边联动信息架构”,设计“动态云边联动机制”,同时结合深度强化学习算法优化系统决策。经试验仿真结果表明,有效地减少终端拜访人员的在途里程数。针对快消品终端拜访场景,能有效地提高公司终端拜访人员的服务水平,维系客情关系,从而达到提高销量、提升订单转化率等目的。

关键词: 终端拜访, 联动决策, 强化学习

Abstract: With the rapid growth of consumer market of FMCG(Fast moving consumer goods), companys' terminal visit cost have increased significantly. At the same time, the terminal visit scene is faced with the problems of obtaining terminal information in real time, making dynamic decisions and responding to customer demand dynamically. The "dynamic cloud-side linkage information architecture" and "dynamic cloud-side linkage mechanism" for FMCG terminal visit problem is proposed. At the same time, the deep reinforcement learning algorithm is used to optimize the system decision. The simulation results show that it can effectively reduce the traveling mileage of terminal visitors. For the FMCG terminal visit scenario, it can effectively improve the terminal service level of enterprises, maintain customer relationship, so as to achieve the purpose of increasing sales volume and order conversion rate.

Key words: terminal visits, linkage decision, reinforcement learning

中图分类号: